@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Science, Faculty of"@en, "Physics and Astronomy, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Hoey, Jesse"@en ; dcterms:issued "2009-01-09T21:03:21Z"@en, "1994"@en ; vivo:relatedDegree "Master of Science - MSc"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """We examine currents and relative zooplankton scattering strengths measured with a vessel-mounted RD Instruments 150 kHz acoustic Doppler current profiler (ADCP) over the southwest Vancouver Island continental margin. Raw ADCP signal strengths are converted into relative zooplankton backscatter intensities using a method of residuals computed from average linear fits to depth profiles, which removes most system-dependent effects. An absolute calibration is not feasible without coincident net tows and a record of system characteristics. The problem of side-lobe contamination of near-bottom zooplankton backscatter estimates is given considerable attention and an objective method is developed to eliminate erroneous backscatter values. Ship-referenced water velocities are converted to absolute (earth-referenced) currents using bottom-tracking measurements of ship velocities. Methods of calculating absolute currents using ship velocities estimated from the change in (GPS) ship position are also examined. It is shown that the GPS positional errors of « ± 300 m lead to velocity errors that are comparable to the currents. Only when bottom-tracking is available are absolute currents reliable. The methods developed to process the data are applied to ADCP data collected in June 1993 over the southwest Vancouver Island continental margin. The currents and relative zooplankton scattering strengths are examined over a 20 hour period along a 14 km survey line which spans the shelf break. Zooplankton concentrations occupied the upper 50 m over the shelf and slope both day and night. Aggregations of scatterers were present over the shelf break at depths of 150-200.m during the day, and migrated to the surface waters at sunset. Mean upwards swimming speeds were 1-1.5 cm/s. Offshore flow and vertical shears of the order of 10⁻³ s⁻¹ on the shelf were correlated with horizontal movements of zooplankton, indicating that the distribution of zooplankton over the shelf and slope was influenced significantly by advection. A strong scattering layer with backscatter intensities 100 times higher than normal surface intensities was advected into the survey area over the shelf a few hours before sunrise on June 30th 1993, and underwent downwards migration just prior to sunrise. Mean downwards swimming speeds were 3-4 cm/s. A southwestward flowing shelf-break current was observed in the upper 50 m with speeds of 15-20 cm/s. The contribution of tidal flows to the net current is determined and is found to be consistent with the predictions of Foreman's (1990) finite-element barotropic tidal model. The ADCP data taken in June, 1993 are examined along eight 100 km long survey lines perpendicular to the coast of Vancouver Island. The eight survey lines span the region from the mouth of Juan de Fuca Strait to the mouth of Barkley Sound. Zooplankton concentrations occupied the upper 50 m over the shelf and slope both day and night. Aggregations of scatterers were present over the shelf break at depths of 150-200 m during the day. A southwestward flowing shelf-break current was observed in the upper 50 m with speeds of 15-20 cm/s. An estuarine type flow was observed at the mouth of Juan de Fuca Strait, with strong (>30 cm/s) outflow in the upper 80 m and weaker landward flow below 80 m. These currents are consistent with previous measurements in the region. The ADCP data taken in June, 1993 are spatially averaged in order to characterize the summer mean zooplankton distribution and current pattern. Surface zooplankton concentrations were observed both day and night over the shelf and slope from Estevan Point to Juan de Fuca Canyon. High concentrations were found near the bottom along the shelf break from Estevan Point to Juan de Fuca Canyon. A southwestward flowing shelf break current with speeds of 15-20 cm/s was persistent along the shelf break from Estevan Point to Juan de Fuca Canyon. Surface outflow from Juan de Fuca Strait was observed to persist up to about 50 km seaward over the shelf. A counterclockwise flowing eddy with a radius of about 30 km was observed over Juan de Fuca Canyon. Current speeds associated with the eddy were strongest near the surface and diminished gradually with depth. Based on the limited data analyzed, it would appear that coastal ocean processes are too affected by high-frequency (tidal period) variability for ADCP surveys to provide synoptic views of the circulation and zooplankton biomass. However, this study demonstrates that the vessel-mounted ADCP is useful for observations of current velocity and zooplankton biomass distribution in confined spatial regions on diurnal time scales."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/3481?expand=metadata"@en ; dcterms:extent "9249254 bytes"@en ; dc:format "application/pdf"@en ; skos:note "I USE OF A V E S S E L - M O U N T E D ACOUSTIC D O P P L E R C U R R E N T P R O F I L E R TO . S T U D Y C U R R E N T S A N D Z O O P L A N K T O N BIOMASS DISTRIBUTION O V E R T H E V A N C O U V E R ISLAND C O N T I N E N T A L M A R G I N By Jesse Hoey B.Sc. Physics, McGi l l University, 1992 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR T H E D E G R E E OF M A S T E R OF SCIENCE in T H E FACULTY OF GRADUATE STUDIES DEPARTMENT OF PHYSICS We accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH COLUMBIA December, 1994 © Jesse Hoey, 1994 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of s\\cs The University of British Columbia Vancouver, Canada Date DE-6 (2/88) Abstract We examine currents and relative zooplankton scattering strengths measured with a vessel-mounted RD Instruments 150 kHz acoustic Doppler current profiler (ADCP) over the southwest Vancouver Island continental margin. Raw A D C P signal strengths are con-verted into relative zooplankton backscatter intensities using a method of residuals com-puted from average linear fits to depth profiles, which removes most system-dependent effects. A n absolute calibration is not feasible without coincident net tows and a record of system characteristics. The problem of side-lobe contamination of near-bottom zoo-plankton backscatter estimates is given considerable attention and an objective method is developed to eliminate erroneous backscatter values. Ship-referenced water velocities are converted to absolute (earth-referenced) currents using bottom-tracking measurements of ship velocities. Methods of calculating absolute currents using ship velocities estimated from the change in (GPS) ship position are also examined. It is shown that the GPS positional errors of « ± 300 m lead to velocity errors that are comparable to the currents. Only when bottom-tracking is available are absolute currents reliable. The methods developed to process the data are applied to A D C P data collected in June 1993 over the southwest Vancouver Island continental margin. The currents and relative zooplankton scattering strengths are examined over a 20 hour period along a 14 km survey line which spans the shelf break. Zooplankton concentrations occupied the upper 50 m over the shelf and slope both day and night. Aggregations of scatterers were present over the shelf break at depths of 150-200.m during the day, and migrated to the surface waters at sunset. Mean upwards swimming speeds were 1-1.5 cm/s. Off-shore flow and vertical shears of the order of 1 0 - 3 s _ 1 on the shelf were correlated with ii horizontal movements of zooplankton, indicating that the distribution of zooplankton over the shelf and slope was influenced significantly by advection. A strong scattering layer with backscatter intensities 100 times higher than normal surface intensities was advected into the survey area over the shelf a few hours before sunrise on June 30th 1993, and underwent downwards migration just prior to sunrise. Mean downwards swimming speeds were 3-4 cm/s. A southwestward flowing shelf-break current was observed in the upper 50 m with speeds of 15-20 cm/s. The contribution of tidal flows to the net current is determined and is found to be consistent with the predictions of Foreman's (1990) finite-element barotropic tidal model. The A D C P data taken in June, 1993 are examined along eight 100 km long survey lines perpendicular to the coast of Vancouver Island. The eight survey lines span the region from the mouth of Juan de Fuca Strait to the mouth of Barkley Sound. Zooplankton concentrations occupied the upper 50 m over the shelf and slope both day and night. Aggregations of scatterers were present over the shelf break at depths of 150-200 m during the day. A southwestward flowing shelf-break current was observed in the upper 50 m with speeds of 15-20 cm/s. An estuarine type flow was observed at the mouth of Juan de Fuca Strait, with strong (>30 cm/s) outflow in the upper 80 m and weaker landward flow below 80 m. These currents are consistent with previous measurements in the region. The A D C P data taken in June, 1993 are spatially averaged in order to characterize the summer mean zooplankton distribution and current pattern. Surface zooplankton concentrations were observed both day and night over the shelf and slope from Estevan Point to Juan de Fuca Canyon. High concentrations were found near the bottom along the shelf break from Estevan Point to Juan de Fuca Canyon. A southwestward flowing shelf break current with speeds of 15-20 cm/s was persistent along the shelf break from Estevan Point to Juan de Fuca Canyon. Surface outflow from Juan de Fuca Strait was iii observed to persist up to about 50 km seaward over the shelf. A counterclockwise flowing eddy with a radius of about 30 km was observed over Juan de Fuca Canyon. Current speeds associated with the eddy were strongest near the surface and diminished gradually with depth. Based on the limited data analyzed, it would appear that coastal ocean processes are too affected by high-frequency (tidal period) variability for A D C P surveys to provide synoptic views of the circulation and zooplankton biomass. However, this study demon-strates that the vessel-mounted A D C P is useful for observations of current velocity and zooplankton biomass distribution in confined spatial regions on diurnal time scales. iv Table of Contents Abs t r ac t i i L i s t of Tables v i i i L i s t of Figures ix Acknowledgements x v i 1 In t roduct ion 1 2 Reg ion of S tudy 5 2.1 Bathymetry 5 2.2 Physical Oceanography 5 2.2.1 Mean Currents 5 2.2.2 Tidal Currents 9 2.3 Biological Oceanography 10 2.3.1 Zooplankton Species 11 2.3.2 Spatial Distributions 12 3 Instrument and D a t a 15 3.1 Acoustic Doppler Current Profiling . . . . . . . . . . . . . . . . . . . . . 15 3.1.1 Transducer Operation 15 3.1.2 Relative Current Measurement 17 3.1.3 Bottom-Tracking 18 v 3.1.4 Absolute Current Measurement 19 3.1.5 Zooplankton Abundance Measurements . . 19 3.1.6 Percent Good Estimate 20 3.2 Internal Data Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 4 Processing Techniques 24 4.1 Data Sets 24 4.2 Bottom Effects 27 4.2.1 A D C P Bottom Depth . . . . . . . . 27 4.2.2 Bottom Scattering: Side Lobes 30 4.2.3 Removing Bottom Scattering Effects for a Constant-Slope Bottom 36 4.2.4 Removing Bottom Scattering Effects for an Irregular Bottom . . . 41 4.3 A G C Calibration 46 4.3.1 Absolute Calibration 46 4.3.2 Residuals 50 4.3.3 Biomass 53 4.4 Absolute Currents 54 4.4.1 Removal of Side-Lobe Scattering Effects 55 4.4.2 Velocity Errors 56 4.4.3 Bottom-Tracking 58 4.4.4 Change in Position 60 4.4.5 Barotropic Tidal Model 67 4.5 Averaging Grids 68 4.5.1 Interpolation Methods 68 4.5.2 Grids 68 4.5.3 Contouring Algorithms 70 vi 4.6 Summary of Procedures 72 5 Resul ts 74 5.1 Mesoscale Cross-Shelf Survey 74 5.1.1 The Survey Line . . 74 5.1.2 Survey Data 78 5.1.3 Backscatter Residuals 81 5.1.4 Currents • • • U5 5.1.5 Zooplankton-Current Correlations 137 5.2 Semi-Synoptic Continental Margin Survey 139 5.2.1 June, 1993 Cruise and Data 141 5.2.2 In-Situ Scattering Layer Distributions 145 5.2.3 In-Situ Current Patterns 150 5.2.4 Spatially Averaged Zooplankton Distributions . . . . . . . . . . . 165 5.2.5 Spatially Averaged Current Patterns . 175 6 S u m m a r y and Recommendat ions 180 Bib l iog raphy 191 A p p e n d i x A Der iva t ion of the Biomass Formula 198 A p p e n d i x B Er rors on Absolute Veloc i ty from B o t t o m Tracking 200 vii List of Tables 4.1 Typical ADCP set-up parameters 26 4.2 Parameters from (4.3 ) and their effect on Sv. 47 4.3 Fitted parameters for the June and October, 1993 cruises in the study area. 52 4.4 Data recorded for each grid point 69 5.1 Eight June vertical sections taken along grid rows and their time and length scales. . 144 5.2 Standard deviations on differences in neighboring profile currents at 10 m. 155 viii List of Figures 2.1 The west coast of Vancouver Island 6 2.2 Theoretical acoustic signatures of several common types of marine organ-isms (Holliday and Pieper, 1980) 12 3.1 Side view of a typical V M A D C P . . • • • • 16 3.2 Filters used in estimating the percent good (RDI, 1989) 21 3.3 A D C P processing cycle (RDI, 1989) 22 4.1 La Perouse Bank C T D lines and profile stations 25 4.2 D T F vs. A D C P estimates of water depth for the June, 1993 cruise. . . . 29 4.3 D T F vs. A D C P estimates of water depth for the October, 1993 cruise. . 3 1 4.4 Typical A D C P transducer beam pattern 32 4.5 Backscattering from a downward-looking A D C P within range of the bottom. 33 4.6 June 1993 time series of pitch and roll angles taken every 2 minutes. Data was recorded from 01:38 to 17:46 P D T on 29/06/93 34 4.7 Six consecutive profiles of raw A G C vs. depth taken on 26/06/93 at 09:46 - 09:58 P D T : 35 4.8 Schematic of three averaging intervals for a ship moving over a constant slope bottom. 37 4.9 Twelve consecutive profiles of raw A G C vs. depth taken on 26/06/93 at 10:48-11:08 P D T . 39 4.9 (continued) 40 ix 4.10 Twelve consecutive profiles of raw AGC vs. depth taken on 26/06/93 at 10:48-11:08 PDT. 43 4.10 (continued) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4.11 Six consecutive profiles of raw AGC vs. depth taken on 29/06/93 at 13:50-14:00 PDT. 45 4.12 Geometric relationships between the ship-referenced and earth-referenced coordinate systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 4.13 Comparison of GPS (top) and bottom-tracking (bottom) estimates of ship velocity for a 6.5 hour period during the June, 1993 cruise from 21:29 PDT on 18/06/93 to 04:07 PDT on 19/06/93. (a) East component. . . . . . . 62 4.13 (b) North component 63 4.14 Differences between GPS and bottom-tracking estimates of ship velocity (top) east and (bottom) north components for the June 1993 cruise. The standard deviations are shown. 64 4.15 Histograms of differences between spatially averaged ship speed'estimates in east and north components for the June, 1993 cruise. The standard deviations are shown. . 66 4.16 ADCP seasonal averaging grid. . 71 4.17 Flow chart for processing ADCP data. . . . . . . . . . . . . . . . . . . . 73 5.1 The ADCP line ship track for June, 1993. 75 5.2 Survey line and local bathymetry for the June, 1993, ADCP survey (Depths in metres). 77 5.3 ADCP survey line profile recording times and distances for the June, 1993 survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 x 5.4 Backscatter residuals (dB) for the June, 1993 A D C P line, (a) Leg 1. Collection times are from 10:08 - 11:52 P D T on 29/06/93 with the ship steaming onshore. 82 5.4 (b) Leg 2. Collection times are from 11:52 - 14:56 P D T on 29/06/93 with the ship steaming offshore. 83 5.4 (c) Leg 3. Collection times are from 14:58 - 16:34 P D T on 29/06/93 with the ship steaming onshore 84 5.4 (d) Leg 4. Collection times are from 16:36 - 19:14 P D T on 29/06/93 with the ship steaming offshore. 85 5.4 (e) Leg 5. Collection times are from 19:16 - 20:54 P D T on 29/06/93 with the ship steaming onshore 86 5.4 (f) Leg 6. Collection times are from 20:56 - 23:24 P D T on 29/06/93 with the ship steaming offshore. Sunset occurred at 21:28 P D T , as shown by the dotted line 87 5.4 (g) Leg 7. Collection times are from 23:26 P D T on 29/06/93 to 01:06 P D T on 30/06/93 with the ship steaming onshore 88 5.4 (h) Leg 8. Collection times are from 01:06 - 03:28 P D T on 30/06/93 with the ship steaming offshore. . . . . . . . . . . . . . . . . . . . . . . . . . . 89 5.4 (i) Leg 9. Collection times are from 03:28 - 05:20 P D T on 30/06/93 with the ship steaming onshore. Sunrise occurred at 5:24 P D T , just after the ship reached the end of the section 90 5.4 (j) Leg 10. Collection times are from 05:20 - 05:50 P D T on 30/06/93. Sunrise occurred at 5:24 PDT, as shown by the dotted line. . . . . . . . . 91 5.5 (a) Time series of backscatter residuals S'v(zk, 4.0 g/m2) of biomass lying in a long narrow (2-3 km wide) filament, the center of which was always within about 5 km from the 200 m isobath. Upwelling along the shelf break contributes to the narrowness of the distribution by effectively 'cornering' the euphausiids between the upwelled water, the ocean bottom at the shelf break and the isolume barrier during the day (Simard et al., 1986; Simard and Mackas, 1989). At night the euphausiids move into the surface waters to feed. The basin aggregations were more spread out and less intense than at the shelf break. High concentrations were observed above Juan de Fuca Canyon, especially in the last week of August, when an upwelling event and a phytoplankton bloom were also noted in this area from salinity, density and chlorophyll profiles. Upwelling of the California Undercurrent waters in Juan de Fuca Canyon can account for the concentrations of plankton seen there, but it is not clear whether this population is resident or experiences large advective gains and losses (Simard and Mackas, 1989). It is also noted that the summer distribution of some of the major vertebrate fishes in the area such as pacific hake coincide almost exactly with the zooplankton distribution, giving an idea of the importance of the zooplankton to the higher trophic levels (Ware and Thomson, 1988). Chapter 3 Instrument and Data This chapter will describe in more detail how the A D C P measures currents, plankton abundances and ship speed from bottom-tracking, followed by a description of the internal data processing that occurs online and the shipboard data collection process. 3.1 Acoustic Doppler Current Profiling 3.1.1 Transducer Operation The instrument used in this study is a narrow-band hull-mounted 153.6 kHz RDI acoustic Doppler current profiler (ADCP) . It was mounted on the hull of the C.S.S. J.P. Tully from the fall of 1992 until the present. The vessel mounted A D C P configuration is shown in Figure 3.1. The A D C P has four transducer heads in a convex Janus configuration with each beam having a width of 4.0° at -3 dB and facing outwards at 30° from the vertical axis of the instrument. The instrument is oriented at 45° from the ship's bow-stern line, giving the four beams azimuthal angles of 45, 135, 225 and 315 degrees with respect to the ship's bow-stern line. The A D C P takes an instantaneous water profile by emitting a pulse (ping) of sound from each of its four beams. Each pulse has a duration of about 3-6 milliseconds, or a length of about 4-8 m. The sound is scattered primarily off zooplankton in the water, and the return echo is range gated into bins of length 2 n (n=l,2,. . .) metres. A short delay between the pulse transmission and the beginning of processing allows the ringing 15 Chapter 3. Instrument and Data 16 sea surface | O O M * / // \\\\ 400 // \\\\ Ral / / ADCP BEAMS \\ \\ meter nge Figure 3.1: Side view of a typical V M A D C P . Two of the four beams are shown looking forward and aft, capable of measuring the along-ship currents and the plankton backscat-tering strength Chapter 3. Instrument and Data 17 in the transducer to die down, and defines a layer below the ship's hull (usually 1-4 m) in which no data is recorded, called the blank-beyond-transmit. The blank-beyond-transmit also ensures that the disturbances in the water caused by the ship's hull are not recorded. The ADCP records the frequency of the returning scattered sound for each bin, from which it determines a vertical profile of the water velocity relative to the ship. It also records the intensity of the return echo for each bin, which gives a vertical profile of the zooplankton acoustic scattering strength. The ADCP has a nominal water profiling range of 400 metres. However, in rough seas, breaking waves generate bubbles which can be entrained under the ship's hull and inhibit the transmission of sound. Bubbles have been known to cause a reduction in profiling range of as much as 60% (RDI, 1989). 3.1.2 Relative Current Measurement At 150 kHz, the primary sound scatterers in the ocean are zooplankton. The plankton have some mobility, but in general the motions of the animals tend to be uncorrelated. That is, they do not tend to school or swim in unison. Due to the large concentrations of plankton in the water, they appear as a group of scatterers passively advected by the local currents (Flagg and Smith, 1989). Thus, the return frequency in each beam will be Doppler shifted according to the motion of the water. The along-beam component of the water velocity, V, can be calculated using the Doppler shift relation: v=\\4. <\"> where Fd is the Doppler shift, Fs is the output frequency, and c is the speed of sound in the water. The ADCP algorithm uses the along-beam velocities from each pair of horizontally opposed beams to calculate one horizontal water velocity component and the vertical water velocity component. This results in two horizontal velocity components in Chapter 3. Instrument and Data 18 the ship's reference frame, (u',v'), and two estimates of the vertical velocity component (w'). Finally, a rotation of the velocity reference frame from ship coordinates (along-ship, cross-ship, z) to earth-referenced coordinates (east, north, z) is made using the pitch, roll and heading angles of the ship determined from navigation devices, giving the velocities relative to the earth, (it, v, w). Inherent in the ADCP algorithm to calculate (u',v',w') is the assumption that the water velocities are the same in each beam for a given bin depth. At a depth of 400 m, beam angles of 30° give rise to a maximum beam separation of 400 m, which sets the horizontal scale over which the assumption of uniformity must hold. Justification of this assumption is found in the error velocity, which is the difference between the two vertical velocity components. A large error velocity indicates that there are substantial differences in the velocities measured by each of the four ADCP beams, either due to transducer errors or because the actual currents are not horizontally homogeneous. The error velocity is a built-in means to estimate data quality (RDI, 1989). 3.1.3 Bottom-Tracking Every second pulse the ADCP emits is designed exclusively to scatter off the bottom. This pulse is used to determine the local water depth and the speed of the ship over the bottom. It is referred to as a bottom-tracking ping and is implemented and processed separately from water profiling pings. It has a longer pulse length and is broken down into as many as 128 depth bins, each shorter than the transmit pulse. The ADCP searches through these depth cells to find the center of the strongest echo, which it chooses as the local water depth. It subsequently uses the measured Doppler shifts of the bottom echoes to estimate the ship's absolute (earth-referenced) velocity. The manufacturers claim that bottom-tracking pings are able to measure ship velocity with an accuracy of about 1 cm/s and bottom depth to within about 1 meter. The bottom tracking range is Chapter 3. Instrument and Data 19 roughly 50% greater than the water profiling range, or 600 m (RDI, 1989). 3.1.4 Absolute Current Measurement As with the water velocity measurements, the estimates of the relative bottom velocity from the bottom-tracking pings are rotated to earth-referenced (east, north) coordinates and recorded for each profile. The absolute, earth-referenced, water velocities can then be computed for each depth bin by a simple subtraction of the bottom velocity from the relative water velocity. If the bottom-tracking algorithm was unable to locate the bottom or was unable to make a reliable estimate of the relative bottom velocity, the ship's velocity must be estimated in some other way in order to compute absolute water velocities. This is typically done using navigational data from the ship's GPS or L O R A N - C system. In absence of any ship velocity estimates, only the vertical shears present in the water profiles can be examined. Most of the effects of the ship's motion can be removed by assuming a reference layer of 'no motion' within the profiling range, and subtracting the velocities measured in the layer of 'no motion' from the rest of the profiles. A l l velocities are then relative to the velocities in the layer of 'no motion', but will not include ship velocity effects. This method reveals horizontal shears on small spatial and temporal scales, while preserving the vertical structure in the water column. 3.1.5 Zooplankton Abundance Measurements Typical input signal levels from scattered sound in the ocean vary by two or three or-ders of magnitude. This input signal must be amplified to a roughly constant level for the signal algorithms to accurately estimate the frequency. This amplification is accom-plished by the Automatic Gain Control circuitry, which controls the gain with a voltage, Vg. The gain-control voltage, Vg, is converted to a logarithmic value and recorded as Chapter 3. Instrument and Data 20 a quantity called the AGC, which reflects the strength and thus the abundance of the sound scatterers. The A G C output is given by: AGC = 1\\og{eIf) (3.2) where 7 , e are constants, Ir is the reflected intensity at the A D C P and Ia is some constant reference intensity. The backscatter intensity, Sv, in decibels is given by: ^ = 10 log( | ) = 101og(g) (3.3) where Ir and Ii are the reflected and incident intensity measured at some reference distance from the scatterer, and as is the total acoustic cross-section. In order to combine the two above equations, the implicit reference distance in (3.3) must be specified. The standard choice is a distance of 1 m, which is not feasible in the case of a ship-mounted A D C P . However, it can be set to the distance from the scatterer to the A D C P by adding the transmission loss of sound through the water due to attenuation and beam spreading. The result is: Sv = kc{AGC) + 20 log R + 2aR + A (3.4) where kc and A are system-dependent constants, R is the ADCP-source distance in metres, and a is the coefficient of sound attenuation in the water in d B / m (Flagg and Smith, 1989). Equations (3.3) and (3.4) demonstrate that the A G C gives a measure of the backscatter strength which can subsequently be related to the acoustic cross-section and ultimately to the zooplankton biomass. This will be discussed in more detail in chapter 4. 3.1.6 Percent Good Estimate The A D C P also makes an estimate of the signal-to-noise (S/N) ratio using a wide and a narrow band filter (Figure 3.2). If the energy passed by the narrow band filter is roughly Chapter 3. Instrument and Data 21 Wide Band Filter Narrow Band Filter • Doppler Shifted Echo. Signal T3 Wide Band Filter Narrow Band Filter Noise -Doppler Shifted Echo Signal Frequency High S /N Ratio ( E n e r g y a EnergyN B) Narrow Band Contains Most of the Energy Frequency Low S/N Ratio (Energy^ = 10 x E n e r g y ) Energy Spread Evenly Over the Wide Band Figure 3.2: Filters used in estimating the percent good. Two extreme cases are shown for high and low S / N ratios (RDI, 1989). equal to that passed by the wide band filter, then the signal must be many times greater than the noise, all its energy being passed by both filters. However, if the signal is small, the energy passed by the narrow filter will be much smaller than that passed by the wide one, since most of the total energy is noise. The S /N ratio is combined with the error velocity mentioned above to produce an estimate of the 'goodness' of the data. This estimate is called the percent good and is recorded for each depth bin. It provides a means of estimating the cut-off depth below which the acoustic data are too weak to yield acceptable results. 3.2 Internal Data Processing Each ping emitted by the A D C P consists of four phases: overhead, transmit pulse, pro-cessing, and sleep (Figure 3.3). The overhead time is used to prepare for ping processing. The pulse is then transmitted, after which there is a short delay to allow for ringing to die down. The A D C P then begins to process the echo. After processing is complete the A D C P goes to sleep, unless it immediately starts another data collection cycle. Each Chapter 3. Instrument and Data 22 SINGLE PING • Processing Time Depends Upon 1) What Processing is Done 2) Number of Cells 3) Speed of Sound, or Computation Time Ping 1 Ping 2 Data Transmission ENSEMBLE OF PINGS Figure 3.3: A D C P processing cycle (RDI, 1989). ping takes about 1-2 seconds to transmit, receive and process, and the A D C P typically emits about 30 water profiling and 30 bottom-tracking pings in a two minute averaging interval. The bin-by-bin averages of the data are then transmitted to the data acquisi-tion system (DAS) to be recorded on disk. The DAS produces records with the following three basic components: Fixed Header This contains information pertinent to the entire session. A session is begun when the DAS is started and is ended when the data acquisition process is interrupted by a user. Typically a cruise will contain one or two sessions. The fixed header contains the A D C P setup, which includes such information as the length of each bin, the number of pings per averaging ensemble, the blank beyond transmit, the number of depth bins, the startup date and time, the cruise ID and various flags which show what parameters were recorded in the data itself. Ensemble Header This contains information pertinent to each ensemble. It includes the recording date and time, the pitch, roll, heading, and temperature readings from ship navigation, the average bottom velocity from bottom-tracking pings, Chapter 3. Instrument and Data 23 the bottom depth, and the latitude and longitude from either GPS or LORAN navigation devices. Ensemble Data This is the information that is transmitted to the DAS from the ADCP, including the full averaged velocity profile in ship-referenced coordinates, the aver-aged AGC profile and the percent good for each depth bin. Chapter 4 Processing Techniques This chapter will describe the basic processing techniques used to convert the raw A G C values into relative backscatter strengths and relative water velocities into absolute (earth-referenced) water velocities. The A G C conversions involve first removing all sec-tions of the A G C profiles which include bottom scattering effects, then calibrating the raw A G C to obtain a clear picture of relative concentrations of zooplankton biomass. This chapter will open with a description of the data used in this study, including the various types of measurements made, the A D C P configuration, and the spatial and tem-poral coverage of the data. It will close with a description of the techniques used to average data temporally onto a regular grid and the algorithms used for the subsequent contouring. 4.1 Data Sets The two different 150 kHz ADCPs used in this study were mounted on the hulls of two ships used extensively in the La Perouse project. There were 24 cruises on the C.S.S. Parizeau: 9 in 1988, 6 in 1989, 4 in 1990 and 5 in 1991. There were 5 cruises on the C.S.S. Tully in 1993. These cruises typically lasted for 1-2 weeks during which the ship spatially covered 30-40% of the area delineated in Figure 4.1. The ship surveys involved primarily conductivity-temperature-depth (CTD) casts, hydro casts and plankton sampling along seven standard survey lines (Figure 4.1), as well as the deployment/recovery/servicing of current meters moored at selected positions. The CTD/Hydro data were initially edited 24 Chapter 4. Processing Techniques 25 30 • 1 2 6 30 1 2 5 Figure 4.1: La Perouse Bank CTD lines and profile stations. Chapter 4. Processing Techniques 26 Sampling Number Bin Blank Mounting Vessel Interval of Length Beyond Depth (sec) depth bins (m) Transmit (m) (m) Parizeau 300 49 8.0 1.0 5.0 Tully 120 128 4.0 4.0 4.0 Table 4.1: Typical A D C P set-up parameters at the Institute of Ocean Sciences. The resulting archived products are described in (Ware and Thomson, 1992). The A D C P was typically set up with the characteristics given in Table 4.1 for each of the two vessels. Sampling was continuous throughout the cruises. The sampling intervals resulted in a spatial resolution at ship speeds of 10 knots of about 1500 nl for the C.S.S. Parizeau and about 500 m for the C.S.S. Tully. A majority of the profiles were taken along the C T D lines, providing extensive spatial coverage in these areas. Other areas were only occasionally covered by the ship when in transit.. Bottom-tracking was turned on for most cruises, giving accurate absolute velocities out to depths of about 450 m, or about 42 km offshore at Estevan Point to about 105 km offshore at the mouth of Juan de Fuca Strait. Most of the data have been edited and processed under contract to Hermes Computing Services (Luc Cuypers). Results include horizontal maps of absolute and relative currents, as well as vertical sections of contoured cross-shore and alongshore components of the water velocities. Contoured vertical sections of the backscatter intensity in decibels also were produced using a calibration technique similar to that described in Chapter 4.3.1. The data used in this study were pre-processed by Luc Cuypers to change the RDI format of the original data files into a more manageable format called PHD300. A n extensive set of analysis programs developed by Luc Cuypers was modified for this study to implement the analysis techniques described in this chapter. Chapter 4. Processing Techniques 27 4.2 Bottom Effects Due to the strength of the bottom echoes compared to that of zooplankton and other scatterers in the water column, sound scattering off the ocean floor is an important factor in both backscatter and velocity profiles when using an A D C P . This section discusses the various effects of the ocean floor on the acoustic echoes: First, the reliability of the A D C P bottom depth obtained from bottom-tracking pings; second, the effects of side-lobe scattering on the water profiles for a flat bottom; and third, the effects of the bathymetry for the cases of a constant-slope bottom and an irregular bathymetry. 4.2.1 A D C P Bottom Depth The bottom depth is obtained from the A D C P bottom-tracking pings, which have a range roughly 50% more than the water profiling range , or about 600 m for a 150 kHz unit (RDI, 1989). The range is slightly smaller over the shelf break and slope since all four beams must be able to find the bottom. This correction is about 25 m for maximum slope angles of 4°, reducing the range to about 575 m. Beyond these depths, the bottom tracking velocity and the bottom depth are flagged and recorded as —19999.0 and —1, respectively. However, in cases where the bottom echo is weak or non-existent, the A D C P has a tendency to interpret any strong scattering layer in the upper 600m as a bottom echo and process the data accordingly, giving rise to a 'false bottom'. In other instances, the A D C P simply misjudges the bottom depth completely. Although the reasons are not yet fully understood, it is assumed that erroneous bottom depths occur because of excessive noise spikes in the bottom tracking data. Because of these problems, the A D C P ' s estimate of the bottom depth cannot be used in the following three cases: 1. The bottom is out of range and the bottom depth is flagged as incorrect; Chapter 4. Processing Techniques 28 2. a false bottom is found; 3. the A D C P 'misjudges' the bottom. A n external estimate of the local water depth is needed to check the reliability of the A D C P ' s estimate. These external bathymetry data were obtained from the Department of Energy, Mines and Resources and are contained in the 'Digital Terrain File' , or D T F , which covers the west coast of Canada with a 1 km resolution grid (Luc Cuypers, person, comm.). A n algorithm is used to find the closest D T F gridpoint to the profile's recorded position and the D T F depth is compared to the A D C P estimate. The A D C P estimate is taken as reliable if one of the following two conditions hold: 1. The A D C P depth is shallower than the bottom-tracking range of 580m and the D T F depth is below the same range following a correction, A / J , for the grid resolution. The D T F grid resolution (±500m) can give rise to maximum depth differences of A D = 500m tan « 35m where « 4° is the maximum bottom slope. Thus, it is required that the A D C P depth is < 580 m and the D T F depth is < 545 m. 2. The D T F and A D C P estimates of water depth are the same to within the resolution of the D T F grid of 35 m depth and the D T F depth is less than an upper limit for the A D C P range of 650 m corrected for the grid resolution, or 650 m — AD = 615 m. Figure 4.2 presents a scatter plot of D T F depths versus A D C P estimates of depth for the June, 1993 cruise together with the limits applied to determine reliable bottom-tracking depths. Most bottom-tracking profiles agree well with the unflagged D T F estimates. A group of points at D T F depths of ~ 750 m have consistently low A D C P estimates of depth, which can not be considered reliable. The October, 1993 cruise showed a similar Chapter 4. Processing Techniques 29 -1 0 200 400 600 800 1000 ADCP bottom depth (m) Figure 4.2: D T F vs. A D C P estimates of water depth for the June, 1993 cruise. The bottom-tracking depths are good inside the solid box. The profiles in which the bot-tom-tracking could not find a bottom have A D C P bottom depth — —1. Chapter 4. Processing Techniques 30 scatter plot (Figure 4.3). The aggregation of points just outside the box at A D C P depths of about 500-550 m are thought to differ slightly more from the D T F estimates due to decreased resolution of the D T F grid past the shelf break. This indicates that condition (2) can be increased to ±60 m without decreasing the reliability of our depth estimates. Henceforth, a profile's water depth will indicate either the A D C P bottom-tracking depth if criterion (1) or (2) was passed, or the D T F bottom depth otherwise. 4.2.2 Bottom Scattering: Side Lobes The detected echo from the bottom is typically about 50 dB higher than the near-bottom water echoes. Due to side-lobe effects, the width of the bottom echo corresponds to about 40% of the water depth . The nature of these side-lobe effects are determined primarily by the A D C P transducer beam pattern (Figure 4.4). The centered main lobe has a beamwidth of 4.0° at the —3 dB level (RDI, 1989), and contains most of the beam power. The remaining power is contained in the sidelobes, which have strengths 30-40 dB below that of the main lobe. The A D C P transducers are sensitive to the two-way beam pattern because they both transmit and receive sound. Thus, the sidelobes are 60-80 dB below the main lobe — a factor of 100 million. Although this level may seem very small, sidelobe contributions become important when the A D C P is measuring signal strengths near the ocean floor. Rocky and sandy ocean floors reflect 100 kHz sound incident at an angle of 30° from normal at roughly the —25 dB level (Urick, 1983). This is 60 dB higher than typical zooplankton aggregation scattering strengths of —85 dB (RDI, 1990). Thus, sidelobe scattering off the bottom ( « —85 to —105 dB) is comparable to main lobe scattering off zooplankton ( « —S5dB), and can therefore contaminate the profile near the bottom. For a downward-looking A D C P with beams angled at 30° from the vertical, the echo from the side lobe facing the bottom will arrive at the transducer at the.same time as that from the main lobe facing 30° from the vertical, so that the Chapter 4. Processing Techniques 31 -1 I , , , -, ; P 1 0 200 400 600 800 1000 ADCP bottom depth (m) Figure 4.3: D T F vs. A D C P estimates of water depth for the October, 1993 cruise. The bottom-tracking depths are good inside the solid box. The profiles in which the bottom-tracking could not find a bottom have A D C P bottom depth = — 1. A clear example of the A D C P finding a false bottom is located at about (20,820). Chapter 4. Processing Techniques 32 o -90 -70 -50 -30 -10 10 30 50 70 90 Direction (degrees) Figure 4.4: Typical A D C P transducer beam pattern. The beamwidth is defined at the -3 dB level. Side lobes are typically located at angles > 15° and are 30-40 dB below the main lobe level (RDI, 1989). bottom 13% [= 1 — cos(30°)] will be contaminated by sidelobes (Figure 4.5). The pitch and roll of the ship also may contribute to the bottom contamination. Absolute values of ship pitch and roll angles taken from the ship's gyroscope are averaged throughout each ensemble interval and recorded for each profile. A 16 hour time series taken from 01:38 to 17:46 P D T on 29/06/93 of averaged absolute pitch and roll angles (Figure 4.6) reveals that roll angles were consistently higher than pitch angles. The maximum roll angles of about 2° will give rise to bottom contamination at the [1 — cos(30° + 2°)] = 15% level. In practise, the width of the bottom echo is slightly larger than 15%. To demonstrate this, six consecutive vertical profiles of raw A G C taken from the June, 1993 cruise are examined (Figure 4.7). During the 12 minutes that these profiles were collected in, the ship was moving over a bottom of constant depth to within one bin length (4 m). The plots show a cut at 85% of the A D C P bottom depth and another at 80%. The 85% cut works well for about two thirds of the profiles, but includes a fraction of the bottom echo for the other third. This 'shoulder' could be a scattering layer sitting within 5 m of the bottom. However, since side-lobe bottom scattering and biological scatters leave Chapter 4. Processing Techniques 33 ADCP Figure 4.5: Backscattering from a downward-looking A D C P within range of the bottom. At this frequency and transducer orientation (30°), the bottom 13% of the profile is contaminated by sidelobe scattering from the bottom. Chapter 4. Processing Techniques 34 o 5 P U5 d o b —i— 10 —i— 15 Time in Hours from 01:38 PDT on 29/06/93 Figure 4.6: June 1993 time series of pitch and roll angles taken every 2 minutes. Data was recorded from 01:38 to 17:46 PDT on 29/06/93. Chapter 4. Processing Techniques 35 Figure 4.7: Six consecutive profiles of raw A G C vs. depth taken on 26/06/93 at 09:46 -09:58 P D T . A D C P estimates of depth from bottom-tracking are shown with a solid line, whereas the 85% cut and 80% cuts are shown with dotted and dashed lines, respectively. Chapter 4. Processing Techniques 36 virtually identical signatures, it is not feasible to try and make the distinction on such a small vertical scale. Thus, the bottom 20% of each water profile is removed in order to get rid of all bottom scattering effects for a flat bottom. 4.2.3 Removing Bottom Scattering Effects for a Constant-Slope Bottom A two-minute averaged profile during which the ship is moving over a flat bottom will be contaminated by bottom scattering at 20% of the minimum depth encountered during the averaging interval. The depth recorded by the A D C P is an average and represents the bottom depth somewhere near the middle of the averaging period. Thus, a bottom sloped at 6 degrees from the horizontal will result in bottom scattering contamination at a percentage of the A D C P bottom depth, D, of: (D - § |V, |Jtan0) x 0.8 1 ~D where Vs is the mean ship speed in m/s, and I is the ensemble averaging interval in seconds. At a ship speed of 10 knots, the maximum Vancouver Island shelf slope angle of 4° causes a change in bottom depth of about 40 m in a 2 minute averaging period. For a minimum shelf depth of 200 m, (4.1) allows contamination at the 28% level, which represents a maximal percentage for the study region. For Vs = 0, (4.1) reduces to the 20% contamination value discussed in the previous subsection. A n implementation of the above formula for computing the effects of a sloped bottom involves making a best estimate of the bottom slope and hence the minimum depth encountered using the bathymetric information available, which consists of the following six bottom depths: 1. Present profile's water depth 2. Next profile's water depth x 100% (4.1) Chapter 4. Processing Techniques 37 Omiin Previous Averaging Interval 2 min Present Averaging Interval 4 min Next Averaging Interval Previous ADCP Depth Previous DTF Depth Present ADCP Depth Present DTF Depth Next ADCP Depth Next DTF Depth Figure 4.8: Schematic of three averaging intervals for a ship moving over a constant slope bottom. The various estimates of depth are shown in their respective positions in time and space. 3. Previous profile's water depth 4. Present profile's independent (DTF) water depth estimate 5. Next profile's independent (DTF) water depth estimate 6. Previous profile's independent (DTF) water depth estimate Figure 4.8 is a schematic of these six depths showing what depth they represent for a ship moving at a constant rate over a bottom with a uniform slope. The D T F depths represent the bottom depth somewhere near the end of the averaging interval as that is when the A D C P queries the ship navigation devices for positional information. Chapter 4. Processing Techniques 38 Therefore, assuming the bottom has a uniform slope, the profile can be cut at 20% of the minimum of the two endpoints in the averaging interval. These endpoints can be found either by averaging water depths, or simply by looking at the D T F depths. Therefore, for a uniform slope bottom, the minimum depth is taken to be the minimum of the following four depths: 1. Average of the present water depth and the next profile's water depth. 2. Average of the present water depth and the previous profile's water depth. 3. The D T F depth for the present profile. 4. The D T F depth for the previous profile. The effect of this depth selection process can be seen by examining a series of 12 consecutive profiles recorded on June 26th, 1993 from 10:48-11:08 P D T , with the ship moving at about 10 knots over a spur canyon to the north and west of the Juan de Fuca canyon at latitude 48.26° N , longitude 125.2° W (Figure 4.9). The cuts at 80% of the A D C P water depth and at 80% of the minimum of the four depths given above are shown as solid and dotted lines, respectively. Over the 20 minute period covered by the data segment the bottom depth increased from about 120 m to about 220 m and then decreased to 100 m. The largest jumps, AH, in bottom depth were between profiles 1324-1325 of AH = 44 m and between profiles 1329-1330 of AH = —43 m. Although the average bottom depth for profile 1325 was close to 180 m, the depth was clearly shallower than this at the beginning of the averaging interval, resulting in the apparent 'scattering layer' situated at a depth of about 120 m. Further evidence that the 120 m scattering layer in profile 1325 was not biological is found in the next three profiles (1326-1328), where this layer was no longer visible. Profile 1329 also shows a similar amount of bottom backscatter influence which occurred when the Chapter 4. Processing Techniques 39 50 Raw AGC 100 150 200 250 50 Raw AGC 100 150 200 250 50 Raw AGC 100 150 200 250 1322 100 200 300 400 1323 1324 400 50 Raw AGC 100 150 200 250 50 Raw AGC 100 150 200 250 50 Raw AGC 100 150 200 250 1325 1326 1327 100 200 300 400 Figure 4.9: Twelve consecutive profiles of raw A G C vs. depth taken on 26/06/93 at 10:48-11:08 P D T Ship speed was about 10 knots. The solid lines show the cut at 80% of the A D C P water depth, the dotted lines show 80% of the minimum of four depths as explained in the text. -Chapter 4. Processing Techniques 40 Figure 4.9: (continued) Chapter 4. Processing Techniques 41 ship was moving from a depth of 164 m to 121 m. Both of these suspect profiles admit the possibility of bottom scattering accounting for the extra scattering layer observed, and must be cut so as to remove the contaminated bins. Figure 4.9 shows that the assumption of a uniform slope bottom is not validated, as the cut developed above does not remove all the bottom scattering effects from profile 1325, leaving a substantial shoulder which is not apparent in the profiles that followed. However, the cut removes nearly all of the shoulder in profile 1329, which can be attributed to the bottom slope being more uniform over the averaging interval of profile 1329 than over the averaging interval of profile 1325. Irregularities in the bottom slope can lead to bottom scattering effects which are not accounted for by a uniform slope model, and so a stricter cut must be devised. 4.2.4 R e m o v i n g B o t t o m Scattering Effects for an Irregular B o t t o m The analysis of the profile data indicates that the Only reliable way to remove all bottom scattering effects from A D C P data is to cut on 80% of the minimum of the following five depth estimates: 1. Present profile's water depth 2. Next profile's water depth 3. Previous profile's water depth 4. Present profile's independent (DTF) water depth estimate 5. Next profile's independent (DTF) water depth estimate This cut will successfully account for most irregular bottom topography occurring on length scales of about 500 m. It will not account for 'bumps' in the ocean floor which have horizontal length scales smaller than about 500 m and vertical scales greater than Chapter 4. Processing Techniques 42 4 m. The effects of this cut are examined on the same twelve consecutive profiles as in Figure 4.9 (Figure 4.10). The dotted lines now represent the above objective cut using the minimum of the five possible depths. The bottom scattering effects in profiles 1325 and 1329 are fully removed by this cut. The five-depth cut overestimates the amount of bottom scattering present in the data, which can become a problem when trying to observe near-bottom biological scatterers. However, as pointed out previously, the acoustic signatures of such scatterers and side-lobe bottom scattering are too similar to be distinguished regardless of the processing approach. Thus, if there is a possibility of bottom scattering effects in the dynamic transducer/ocean-floor configuration, then elimination of these effects must take prece-dence over the possibility of biological scatterers to avoid any ambiguities in the final results. For example, consider six consecutive profiles taken on June 29th, 1993 from 13:50-14:00 P D T at latitude 48.05° N , longitude 125.6° W on the shelf break near the southern entrance to the Juan de Fuca Canyon (Figure 4.11). The bottom increases from a depth of about 140 m at profile 1414 to a depth of about 280 m at profile 1417, with the biggest jump (AH = 52 m) between profiles 1414-1415. The effect of this 52 m change in bottom depth is manifested in profile 1415 as an apparent 'scattering layer' similar to what was seen in the profiles of Figure 4.9. However, in this case, the successive profiles also show a similar structure lying about 100 m above the bottom. This would seem to indicate that the 'scattering layer' in profile 1415 is due, at least in part, to biological scattering. It still must be removed because it is possibly due in part to bottom scat-tering of acoustic side-lobes from the shallower depths recorded near the beginning of the averaging interval. The corresponding scattering layers in profiles 1416-1420 are not removed by the cut, and so the presence of this biological feature is not altogether lost. The most efficient method for removing all bottom scattering effects would be to record the minimum depth encountered in each averaging interval. Chapter 4. Processing Techniques 43 50 Raw AGC 100 150 200 250 50 Raw AGC 100 150 200 250 100 200 300 400 1322 100 200 300 400 1323 100 200 300 400 Raw AGC 50 100 150 200 250 1324 50 Raw AGC 100 150 200 250 50 Raw AGC 100 150 200 250 50 Raw AGC 100 150 200 250 100 200 300 400 1325 100 200 300 400 1326 100 200 300 400 1327 Figure 4.10: Twelve consecutive profiles of raw A G C vs. depth taken on 26/06/93 at 10:48-11:08 P D T . Ship speed was about 10 knots. The solid lines show the cut at 80% of the A D C P water depth, the dotted lines show 80% of the minimum of five depths as explained in the text. Chapter 4. Processing Techniques 44 Figure 4.10: (continued) Chapter 4. Processing Techniques 45 100 200 300 400 Raw AGC 50 100 150 200 250 1413 50 Raw AGC 100 150 200 250 50 Raw AGC 100 150 200 250 100 200 300 400 1414 100 200 300 400 1415 50 Raw AGC 100 150 200 250 50 Raw AGC 100 150 200 250 50 Raw AGC 100 150 200 250 1416 100 200 300 400 1417 1418 100 o o T3 3-200 300 400 400 Figure 4.11: Six consecutive profiles of raw A G C vs. depth taken on 29/06/93 at 13:50-14:00 P D T . Ship speed was about 10 knots. The solid lines show the cut at 80% of the A D C P water depth, the dotted lines show 80% of the minimum of five depths as explained in the text. Chapter 4. Processing Techniques 46 .4.3 A G C Calibration The relation between the raw A G C and the backscatter strength, Sv, was given by (3.4): Sv = kc(AGC) + 20 log R + 2aR + A (4.2) The A G C can be calibrated, to give absolute backscatter strength and thus the acoustic cross-section by estimating the constants kc and A. In absence of such an estimate, residuals can be calculated from fits of the vertical profiles of A G C vs. 20 log R + 2aR, yielding relative acoustic backscatter only. These two methods are discussed in the following sections. 4.3.1 Absolute Calibration The unknown constants kc and A in (4.2) are functions of system dependent parameters, a breakdown of which is obtained by expanding the sonar equation using the system characteristics (RDI, 1990; Urick, 1983): Sv• = 10log(4.47 x l O - 2 0 ^ ^ ( ^ + 273)) . (4.3) , H-101og( l f J f c «^- . ) / 1 0 . - l ) + 20logR + 2aR /. -10.1og(cP)_-10iog^i where k2 = 4.3, ks .= 4.17 x 105, Tx = temperature of transducer head (°C), kc = 127.3/(7; +273), -Te — temperature of electronics (°C), E = A G C in counts, Chapter 4. Processing Techniques 47 Parameter Depends on Range ASV Value Tx water T° [5,15]°C 0.15dB 10° kc electronics T° [5,20]°C 3dB 18° Er noise [25,35]dB . 4dB 30dB a depth [0,400]m l d B 0.044 a water T° [5,15]°C 0.5dB 0.044 c sound speed [1460,1530]m/s 0.2dB 1475.1 m/s Vs Voltage [219,221] V 3dB 220 V Table 4.2: Parameters from (4.3) and their effect on Sv. Values given in the last column are those used in this study. Er — Reference level of echo intensity (electronics noise) in counts, R = Along-beam distance from transducer to depth bin (m), a = attenuation coefficient (dB/m), c = speed of sound in water (m/s), P = Pulse length (m) (typically P = bin length), K1= Power into the water = [^ x n ^~ 6 ] 2 x Klc (W), V, = V A C input, a = 1.397, b = 4.27, n = 37.14, Klc =3.3 (W) ; The values of k2, ks, a, b, n, and K\\c in. (4.3) are nominal values given in (RDI, 1990), and contribute errors of less than ±2dB to the estimation of Sv (RDI, 1990). The remaining six parameters: Tx, kc, Er, a, c and Vs must be measured independently for each instrument. Table 4.2 shows these six parameters, what they depend upon, how Strongly typical variations in the dependencies will affect an estimate of Sv using constant parameters, and a best estimate of the values. * Chapter 4. Processing Techniques 48 The temperature of the transducer head, Tx, can be taken as the temperature of the water under the ship's hull. Due to its small effect on Sv, Tx can be set to a constant value of 10°C, a characteristic temperature for the outer coast near 5 m depth. The temperature of the electronics, Te, on the other hand, can have a substantial effect on Sv. A small change in Te, 8Te, causes a change in Sv, 8SV, of: 8SV = E8kc = E dkc dTe 197 ^ 8Te = E ,„ 1 X 9 8Te (4.4) (Te + 273)2 v ' For characteristic values of E = 200 counts and Te = 18°, (4.4) gives 8SV = 0.3 8Te. Thus, to reduce 8SV below 2 dB, the temperature of the electronics unit must be kept at constant temperature to within about ±3°C. The unit which houses the A D C P electronics was located in the lower laboratory on the C S.S. Tully, which is generally heated to a uniform temperature. A 8Te < 6°C is assumed for the data used in this study. Only a consistent record of the temperature near the electronics unit could reduce the error in Sv induced by 8Te further. Changes in the speed of sound affect the estimation of Sv very minimally, and can be ignored. The A D C P uses a value of 1475.1 m/s for depth cell size calculations, and this value is adopted for the calculations in this analysis. The attenuation coefficient, a, changes with the temperature of the water, and so will be variable above the thermocline and roughly constant below it, where the water temperature is more homogeneous. A variation of ±2°C in the deep water causes a change in a of less than 0.001 dB/m. Thus, a specified nominal value of a for seawater at 4°C and 35ppt of 0.044 d B / m (Urick, 1983; RDI, 1990) is good to within about 2% (0.001 dB/m) in the deep, isothermal water. The nominal value will be about 0.008 d B / m low in the shallow water. However, at these depths it has little effect on Sv, since the 2aR term increases linearly with depth. Use of the deep-water value of a will therefore induce an error in Sv of less than 0.1 dB at 50 m, increasing to about 0.8 dB at 400 m. Chapter 4. Processing Techniques 49 The A C power to the transducer head most probably underwent a smaller range of variability than the one given in Table 4.2. Consequently, the error in Sv due to variations in K\\ could probably be reduced to below 1 dB with a single measurement of . To reduce this error further would involve monitoring the transducer power for a longer period of time to detect any changes and to correct for them. There are two methods to estimate the background noise in the A D C P electronics, ET (RDI, 1990): (1) The approximation method, in which the value of Erc given by RDI is adjusted for real-time electronics and transducer temperatures, giving Er; (2) The in-situ method, in which the value of Er is obtained from the data set by examining echo intensity values taken from depth bins from which no anomalous echo is present. Since Er is a measure of the thermal noise of the A D C P electronics, care must be taken to insure that there are no other interfering noise sources contributing to the returning echoes. Thus, these measurements should be taken while the ship is motionless on a calm sea with engines off. The range of the profile and the bottom depth must be large enough so that propagation losses reduce the signal-to-noise ratio well below unity. At this point, the recorded value of the A G C is the noise level, Er. Heywood et al. (1991) chose to estimate Er when the percent good parameter had fallen to zero. A similar analysis was used for the October, 1993 cruise. The result was a noise level of 30 dB ± 5 dB. A specific calibration run would be necessary to obtain this value to any degree of accuracy. RDI recommends a separate calibration of ETC before each experiment and suggest the in situ method because the A D C P is measuring the reference value in real-time (RDI, 1990). The summed errors on all the parameter estimates will result in a total systematic error on Sv of about 5 dB. This yields a systematic error on the scattering cross-sections calculated from Sv of roughly the same magnitude as the cross-sections. Therefore, more precise calibration of Sv is needed to make this approach useful. Chapter 4. Processing Techniques 50 4.3.2 Residuals The computation of residuals presented in this section offers an alternative to the abso-lute calibration method described in this previous section. Although it makes no claim at being able to determine absolute zooplankton scattering strengths, it gives a good estimate of the relative structure. The proposed residual method makes the assumption that the backscatter strength, Sv, consists of a constant background level, S*, plus a. small depth-dependent perturbation or residual, S'V(R): SV(R) = S'v + S'V(R). (4.5) Equation (4.2) then becomes: E = ~{20logR + 2aR) + hs:-A) + ±-S'v. (4.6) kc and A are constant (in R) since they depend only on characteristics of the transducer and the electronics. Thus, a linear fit of E vs. the transmission losses, (20 log R + 2aR), will result in a slope and intercept of: dE 1 1 R dS'v P G ~ d(20 log R + 2aR) ~ kc + kc (20 + 2aR) dR ' { } 1 , 1 , intercept = £ | ( 201ogf l+2aR=0) = T-^v ~ A) + 7-(£0l(201ogJH2c«rt=0)- (4.8) Kc Kc Since (20 log R + 2aR) = 0 at R « 0.99, which is less than the blank-beyond-transmit, the value of (S'v)\\(2ologR+2aR=o) is arbitrary and is chosen to be (5(,)|(2oiogJM-2aii=o) = 0, leaving the intercept = —^-(S* — A). If the backscatter residuals are uniform, such that: ^ ( 2 0 ^ ( 4 9 ) then the slope ~ — which, together with the intercept (= —^{S* — A)), yields values of kc and (S* — A), enabling calculation of the residuals using (4.6): S'v = kcE + 20 log R + 2aR- (S*v - A) (4.10) Chapter 4. Processing Techniques 51 The residuals calculated in (4.10) give the vertical structure of zooplankton scatter-ing strengths for each profile separately. In order to examine horizontal structures or to compare two profiles, the residuals must be computed from the same fit parameters, which are chosen as the mean fit parameters for the entire data set in question. Such an averaging procedure relies on the assumption that S*, A, and kc remain constant throughout the study period. If this assumption holds, the fitting procedure will effec-tively remove all the system-dependent components from the backscatter profiles, leaving a cruise-independent set of backscatter strength estimates. There were two assumptions made in the preceding development which are now dis-cussed. To verify the first assumption (4.9) , the average fit parameters (kc and S* — A) are calculated using all the profiles taken on the Vancouver Island continental margin dur-ing two 1993 cruises (Table 4.3). The profiles are broken down into five groups depending on their bottom depths. A l l profiles are included in the first group (bottom depth> 0 m), and only those taken over bottom depths >320 m are included in the last group (bottom depth> 320 m). The mean fit parameters (kc and S* — A) for profiles taken over the slope only (bottom depth>240 m) have standard deviations of about 10% (Table 4.3). If the profiles taken over the shelf (bottom depth<240 m) are included in the averages in Table 4.3, the standard deviations increase to 30-50%. This increase indicates that the scattering distributions over the slope are more uniform (more in accordance with (4.9)) than over the shelf. The fits to profiles taken over the slope are therefore more represen-tative of the transmission losses. Fitted values of kc and S* — A for profiles taken over the slope only are assumed to approximate the actual values more reliably than the fitted values for all profiles together. Physical oceanographic processes over the shelf and shelf break appear to be responsible for more localized high concentrations of scatterers in the shelf and shelf-break waters which lead to the higher variances in the fitted parameters observed. The standard deviations in Table 4.3 indicate that the best estimates of kc Chapter 4. Processing Techniques 52 kc s: - A cruise bottom depth > mean std.dev. mean std.dev. 0 0.45 0.19 103.0 26.0 9311 80 0.47 0.20 104.4 27.4 June 160 0.53 0.07 109.3 5.73 1993 240 0.54 0.06 110.3 4.62 320 0.54 0.06 110.1 4.32 0 0.41 0.14 92.7 20.2 9335 80 0.45 0.15 98.9 20.4 October 160 0.58 0.12 115.6 13.2 1993 240 0.57 0.08 114.1 6.9 320 0.56 0.06 113.2 4.0 Table 4.3: Fitted parameters for the June and October, 1993 cruises in the study area. Profiles with bottom depths greater than the value in the first column (bottom depth >) are included in the averages. A value of bottom depth >0 indicates that all profiles were used in the average. and S* — A were obtained from fits to profiles with bottom depths >320 m, and hence averages of only fit parameters from fits to profiles taken over the slope are used in the residual calculations. The second assumption was that 5*, A, and kc were constant throughout the study period. Following the analysis of Chapter 4.2.1, the only parameter which could seriously weaken this assumption is the electronics noise, Er. A l l others are stable enough to not affect Sv by more than about 2 dB. The electronics noise undergoes, changes primarily due to the temperature of the electronics. The effect of temperature on Er is roughly given by (RDI, 1990): Er = constant - 0.9Te - 0.LT x (4.11) Thus, the changes in Te and Tx described above will result in a change in Er of about 1 dB. The ranges of the dependencies in Table 4.2 can be used to calculate the variability in kc and A which are expected during either cruise: okc = 0.005 dB/m (4.12) Chapter 4. Processing Techniques 53 oA = 2dB. (4.13) Thus, the assumption that S*, A, and kc are constant holds to within about 2%. The variances in Table 4.3 are due to a combination of the effects of (4.9) being violated too strongly and of changing system characteristics (as in Table 4.2). Increased violation of (4.9) only yields greater variances in the fit parameters. However, this type sample average used to obtain the fit parameters for the June, 1993 cruise in Table 4.3, the error in kc and S* — A due to violation of (4.9) is reduced to under 0.1%. The bias errors in kc and A due to changing system characteristics, on the other hand, cannot be reduced with increased sampling, are given by (4.12) and (4.13). Using (4.10) and a representative value of E=200 counts, the errors in Sv due to o~kc and OA are found to be roughly 2 dB. Therefore, the values of the average fit parameters are accurate enough to remove all system-dependent effects from the backscatter residuals to within about 2 dB. 4.3.3 Biomass Zooplankton biomass can be calculated by first making estimates of the relative abun-dances of the various species of zooplankton in the water, and then specifying a scattering model for each species. A simple calculation starting from (3.3) gives the total biomass per unit volume, W (see appendix A): of error can be reduced by including many samples in the averaged fits. For the 10000-(4.14) where W 5, V Biomass per unit volume (g/m3), Acoustic backscatter per unit volume (dB) relative number abundance of species i, r-T' where n,- is the number of plankton of species i 2 v i Chapter 4. Processing Techniques 54 W{ = average weight of individual member of species i ( « 4°, causing systematic errors in water ve-locities of 8 cm/s due to scaling bias and 36 cm/s due to misalignment angle (for a 10. knot ship speed). Scaling bias for the bottom velocity was found to be about 0.3% of ship speed which resulted in a systematic error in water velocities of 1.5 cm/s. Ramp et al. (1991), also using a 150 kHz vessel-mounted A D C P , found a scaling bias of 0.3% and a misalignment angle of only 1°, leading to errors of about 10 cm/s in water velocities. The misalignment angle for the A D C P being used in this study has never been measured with any precision. Ship pitch and roll typically introduces errors of ~1 cm/s in the horizontal currents (Kosro, 1985; RDI, 1989). Chapter 4. Processing Techniques 58 4.4.3 Bottom-Tracking A bottom-tracking ping is emitted for each water profiling ping, and the resulting bottom velocities are converted to earth-referenced coordinates (east, north), which can then be subtracted from the relative, water velocities to obtain the absolute water velocities. Random error in the bottom-tracking velocity is quoted by R D I as being co a. a. o o o o CM O o CM O O T £ 5 o a (A E f CO o o o o CM O o CM o o t (b) 3 4 Time (hours) 3 4 5 Time (hours) Figure 4.13: (b) North component Chapter 4. Processing Techniques 64 Figure 4.14: Differences between GPS and bottom-tracking estimates of ship velocity (top) east and (bottom) north components for the June 1993 cruise. The standard deviations are shown. Chapter 4. Processing Techniques 65 standard deviations in the mean ship speed differences shown in Figure 4.14 include er-rors from both bottom-tracking and GPS estimates of ship velocity. However, the errors in the bottom-tracking estimates were of order 1 cm/s, and so the errors in the GPS estimates of ship velocity alone, (CTU)GPS, (CV)GPSI were roughly: The errors in absolute water velocities will be higher than those in (4.26) and (4.27), due primarily to the differences between the heading estimates from the relative water velocities, 0, and from the GPS-derived ship velocities, 0n. In any case, the resulting absolute water velocity errors will greatly exceed the typical currents of 10-20 cm/s. As a consequence, individual GPS estimates provide unacceptable surrogates of bottom-tracking velocity. The spatial averaging onto a grid substantially reduces the variability in the GPS velocities. Since at least 10-15 water profiles are averaged onto each initialized grid point, the standard deviations should be reduced by a factor of 3-4. The differences in GPS and bottom-tracking estimates of spatially averaged ship speeds are presented in histogram form in Figure 4.15. Again, the standard deviations in Figure 4.15 include errors from both bottom-tracking and GPS estimates of ship velocity. However, the errors in the spatially averaged bottom tracking estimates are also a factor of 3 to 4 smaller (<0.5 cm/s), and so the errors in the spatially averaged GPS estimates of ship velocity are roughly 10-12 cm/s. Such errors represent a minimum value obtainable on the spatially averaged absolute water velocities, due to heading differences as described above. The minimum errors in spatially averaged water velocities are still of the same magnitude as the typical flows to be measured, indicating that only very strong (^ > 10 cm/s), large-scale water flow patterns will be resolved when using spatially averaged GPS estimates of ( bOcm/s) occur (Figure 4.15) and will dominate the water velocities. These outliers typically are the result of only one or two profiles being averaged onto a grid point and could be eliminated by requiring a certain number of profiles for each included grid point. 4.4.5 Barotropic T ida l Model A finite-element, barotropic tidal model developed by Foreman and Walters (1990) is used to give a rough estimate of the magnitude of the tidal currents in the study area. The model uses a triangular grid with 2050 nodes and 3597 elements covering most of the area of interest to this study, extending from Juan de Fuca Strait to the southern tip of Estevan Point and from the coast to a distance of about 150 km offshore. Since the baroclinic tidal effects have small sea surface manifestations, such a barotropic model can reproduce the surface elevations accurately. However, the model cannot account for the baroclinic tidal currents in the region. The tides predicted by the model were compared with extensive tide-gauge and cur-rent meter observations, and expected accuracies in the amplitudes were obtained (Fore-man and Walters, 1990). The tidal currents did not compare as favorably as the tidal elevations, although the differences in the current ellipse parameters between the model and the observations were only around 1-2 cm/s (major and minor semi-axes) and 5° —15° (phase lag). Significant differences with depth were also observed in the M?, component, and are due primarily to the combined effects of a barotropic Kelvin wave and an inter-nal tide. The predicted tides were also compared with Flather's finite-difference model (Foreman and Walters, 1990; Flather, 1987), and both predicted high tidal currents (10-20 cm/s) close to the mouth of the Juan de Fuca Strait, decreasing north and west to about 2-5 cm/s along the shelf break. The barotropic model will only be used in this thesis as a qualitative measure of the Chapter 4. Processing Techniques 68 large scale tidal effects. The circulation patterns discussed in chapter 5 will be comprised of both mean and tidal currents, and the model will be used to isolate any major tidal effects which dominate the local mean currents. 4.5 Averaging Grids 4.5.1 Interpolation Methods In order to facilitate contouring of vertical sections and horizontal maps of both currents and backscatter intensity, the data are smoothed (spatially averaged) on a regular grid. This is done simply by locating the nearest grid point to each profile's position and assigning to that grid point the absolute currents and raw A G C . The A G C data are separated into night and day profiles, using code written by Luc Cuypers, adapted from code provided by Chris Aikman of the Dominion Astrophysical Observatory. The sunrise and sunset are both 'upper limb' values, meaning a 0° angle between the top of the sun's disc and the horizon. A record is kept during the averaging process of the number of samples added to each grid point, which is used after all profiles have been processed to average them on each grid point. The data in Table 4.4 are recorded for each grid point. The number of cruises is always 1, unless averaging multiple cruises together. Grid points or depth bins with no data are flagged with -0.1E+05. 4.5.2 Grids Two grids are used in this study. The first is a large grid which covers the entire area of interest to this study with a resolution of 3.5 km in the cross-shore direction and 8.0 km in the alongshore direction (Figure 4.16). The grid rows are at angles of 39° to the west-east direction, so as to roughly follow the general trend of the C T D survey lines, Chapter 4. Processing Techniques 69 1. Depth (m) 2. Total residual backscatter intensity (dB) 3. Total time in (2) (min.) 4.. Number of samples in (2) 5. Number of cruises in (2) 6. Night residual backscatter intensity (dB) 7. Total time in (6) (min.) 8.. Number of samples in (6) 9. Number of cruises in (6) 10. Day residual backscatter intensity (dB) 11. Total time in (10) (min.) 12. Number of samples in (10) 13. Number of cruises in (10) 14. u absolute (cm/s) 15. v absolute (cm/s) 16. Total time in (14),(15) (min.) 17. Number of samples in (14),(15) 18. Number of cruises in (14),(15) 19. Percent good for (14),(15) 20. u relative (cm/s) 21. v relative (cm/s) 22. Total time in (20),(21) (min.) 23. Number of samples in (20),(21) 24. Number of cruises in (20),(21) 25. Percent good for (20),(21) Table 4.4: Data recorded for each grid point. Chapter 4. Processing Techniques 70 which lie in a cross-shore direction. The second grid is a smaller one used for the June 1993 A D C P line data, and is comprised of 2 columns and 64 rows, with a gridpoint spacing of about 1.8 km by 0.5 km. Most of the data is centered along a line which runs close (within about 0.3 km) to one of the grid columns. 4.5.3 Contouring Algorithms Two interpolation algorithms are used in the analysis. Both define a regular grid onto which the data is interpolated. The gridded data.is then contoured. The first is an inverse distance method, in which the data is weighted according to the inverse of the square of the distance between data point and grid point. The region of influence of grid points is two grid units, meaning that a data point value influences grid points two grid units away from it (Cuypers, 1993). The second is a linear method, where the surface is approximated by triangular patches joining groups of three data points, and linear interpolation is used within the triangle. (Stat. Sci., 1991; Becker et al., 1988). Most of the data used in this study was averaged onto a regular grid prior to the use of interpolation and contouring algorithms, so that the interpolation grid size used was that of the original grid. In the cases where the data was not gridded prior to interpolation, an interpolation grid size was chosen which reflected the distribution of the data. For this type of previously un-gridded plot, the raw data was examined to ensure that the grid size was appropriate, and that the contours did not cause any spurious elements to appear in the surface. The inverse distance method is used for all plots done with the Oceanplot1 software package. This package produces contours which are rounded using two arcs per straight line segment, resulting in few sharp edges in the contours. The linear method is used for 1 Oceanplot Version 1.01, ©1993 Hermes Computing Services Chapter 4. Processing Techniques 71 Figure 4.16: ADCP seasonal averaging grid. The grid spacing is 3.5 x 8.0km and the grid rows are angled at 39° counterclockwise to the eastward direction. Chapter 4. Processing Techniques 72 all plots done with the S-Plus 2 plotting software. This package produces straight-line contours with no smoothing, usually resulting in sharp edges at each grid point. 4.6 Summary of Procedures Figure 4.17 shows the data processing procedure, starting with the raw data files (ping-data.***) produced by the A D C P data acquisition system. PhdSOO and gridadcp are Microsoft QuickBasic and FORTRAN programs, respectively, written by Luc Cuypers and modified by the author to implement the processing techniques described earlier in this chapter. Newbinrs and newavgrs are also Microsoft FORTRAN programs, written by the author, which fit the raw A G C profiles to determine the average fit parameters, compute the residuals using these average fits, and change the vertical structure from 4 m to 8 m bins. Velout is also a Microsoft FORTRAN, which is used to produce ASCII listings of the fitted data files for use with the S-Plus 3 plotting software. Oceanplot 4 is a software package used to produce vertical sections and horizontal maps. 2S-Plus version 3.0, © 1991, Statistical Sciences, Inc. 3S-Plus version 3.0, © 1991, Statistical Sciences, Inc. 4 Oceanplot Version 1.01, ©1993 Hermes Computing Services Chapter 4. Processing Techniques Raw Data Files (plngdata.***) phd300 Velocity cut at 50% 1 Agcprofs.***, Curprofs.*** gridadcp Cut AGC profiles at 80% of shallowest of five depths as described in section 4.2 Cut current profiles at percent good = 50% Absolute currents computed from bottom tracking whenever posssible Otherwise from ship navigation (GPS or LORAN-C) 193»*grid with raw AGC | newbinrs newavgrs Average fit parameters, TT„(S; - A) computed from fits to profiles with depth > 320m 93**fit with average fit residuals velout OCEANPLOT Horizontal maps and vertical sections of absolute currents and backscatter residuals. Figure 4.17: Flow chart for processing A D C P data. Italicized names are Fortran QuickBasic programs, boldface names are data files. Chapter 5 Results This chapter describes results obtained from A D C P data collected during the June, 1993 La Perouse cruise from 19/06/93 to 07/07/93, using the processing techniques developed in the last chapter. Chapter 5.1 examines a subset of the data which covers a small spatial area (14 km by 1.5 km) and has a short (19 h) duration. The basic zooplankton distributions as well as the current patterns are described. Correlations between the zooplankton distributions and the current patterns are determined. Chapter 5.2 describes all the data taken during the June, 1993 cruise over the southwest Vancouver Island continental margin. The data are first examined insitu on temporal scales of about 10 hours and spatial scales of 100 km cross-shore and 20 km alongshore. Then, the data are examined as an average over the entire cruise period and over spatial scales of 100 km cross-shore and 200 km alongshore. A l l times referred to in this chapter are in P D T . 5.1 Mesoscale Cross-Shelf Survey .5.1.1 The Survey Line During the June, 1993, La Perouse cruise, a 19-hour period was devoted exclusively to steaming back and forth over a survey line to gather a time series for a small area over the Vancouver Island shelf break and slope. Figure 5.1 is a map of the ship's track during this period. The ship started at the southwest corner (48.22° N , 125.80° W) , on June 26, 1993 at 10:10, and proceeded to steam the survey line ten times. Onshore steaming \" . ' ' '74 • Chapter 5. Results 75 a> i , : , . , , : , 1 -125.82 -125.80 -125.78 -125.76 -125.74 , Longitude Figure 5.1: The ADCP line ship track for June, 1993. The ship started at the southwest corner at 10:10 on 29/06/93, and left from the northeast corner at 5:54 on 30/06/93. The dots denote averaging grid centers. The four CTD stations B11C, LB11, B11B and B11A are shown as solid triangles. Chapter 5. Results 76 was at a constant ship speed of around 5 knots, and required about 1.0-1.5 hours for each leg. Offshore steaming included 10-20 minute stops for C T D casts at four standard C T D stations labeled B11C, LB11, B11B, and B11A (Figure 5.1). Steaming between C T D stations was at ship speeds of about 5 knots. Each offshore leg of the A D C P line took about 2.0-2.5 hours. Upper limb sunset occurred at 21:28 on 29/06/93 , during the 6th leg of the A D C P line. The program continued throughout the night with four more passes over the line. Upper limb sunrise occurred at 05:24 on 30/06/93, just before the ship left the area from the northeast corner (48.33° N , 125.75° W) at 05:54 P D T . The first 9 legs covered about the same 14 km, of the shelf-slope region, whereas the 10th leg only covered the northeasternmost 3 km of the line. A total of 591 vertical profiles of currents and A G C were recorded at 2 minute intervals. . -The local bathymetry determined from D T F depths is shown in Figure 5.2. The survey line was located along the northwest side of a small canyon which penetrates about 10 km into the shelf in a northeastward direction. The survey line was roughly perpendicular to the local isobaths, although with an angle of 70° from the east, it was somewhat askew to the general direction of the shelf isobaths, which run perpendicular to about 40° from the east. The deepest water encountered was just over 600 m at the southwest corner of the survey line, the shallowest about 180 m at the northeast corner. Most A D C P bottom depths from bottom-tracking were reliable, with only about 5% of the depths rejected by the criteria presented in Chapter 4.2.4. Rejected depths were always in the far southwest corner of the line, and primarily during the last 4 legs of the survey session. Since the ship was usually turning when sampling over the deepest water, the areas where the GPS estimates of ship velocity were used in place of the bottom-tracking velocities have unreliable absolute water velocities, and hence are not included in the current analysis. Chapter 5. Results 77 Figure 5.2: Survey line and local bathymetry for the June, 1993, A D C P survey (Depths in metres). Chapter 5. Results 78 5.1.2 Survey Data The ADCP data collection during the survey period provided a time series of vertical sections along the survey line. The current and backscatter data are functions of the form: V(d, t,Zk) and S'v(d,t,Zk), where d is the distance from the southwest corner of —* the survey line, t is the time, and Zk is the depth of bin k. V and S'v are defined at a series of irregular (in d and t) and regular (in z) points in the three dimensional space (d,t,z). The times, t, for each of the 591 profiles in the survey are plotted in Figure 5.3, as a function of the distance, d. The cruise legs are labeled 1 through 10, each of which —* is nearly an instantaneous picture of V and S'v over the vertical section d versus Zk for t « constant (to within 1-2 hours). There are two other types of sections that can be generated to visualize the four-dimensional nature of this time series. The first are plots of V and S'v as functions of Zk and t only [V^z*, t) and S'v(zk, t) for fixed d], which reveal the time evolution of the vertical structure for each point along the survey line. The second are plots of V and S'v as functions of d and t only, [V(d, t) and S'v(d,t) for fixed Zk], which reveal the time evolution of the horizontal structure for each depth. The grid in Figure 5.1 was used for the plots of (d, Zk) and (zk, t) to average the data as described in Chapter 4.5. The northernmost grid row is used for most data points; the southern grid row is used only for short periods when the ship drifted off the main survey line. Only the northernmost grid row is used in the following analysis. The averaging process results in current and backscatter values defined at regular points in all three dimensions, or functions of the form V(d{, tj, Zk) and S'v(d{,tj, Zk), where d{ is the distance from the SW corner to grid-point i, (i =1 to 32), Zk is the depth of bin k, and tj = t(d{,j) is an array containing a set of 32 PDT times for each of the ten legs of the survey line. Each t(di,j) is calculated as the mean of the recording times of all profiles averaged onto grid-point d{ during cruise leg j. Throughout this analysis, the Chapter 5. Results 79 r 8 5 W E o b o '$ u> Q t— O B11C 1500 1600 1700 1800 1900 2000 2100 Ensemble # Figure 5.3: A D C P survey line profile recording times and distances for the June, 1993 survey. C T D stations and the leg numbers (1-10) are displayed. Times are P D T . Chapter 5. Results 80 times, t, will usually be given in hours from the beginning of the survey period (10:10 on 29/06/93). The first 9 legs each typically covered about 90% of the northern grid row, leaving out the top and bottom two grid-points. Each grid-point usually contains an average of 1-3 profiles, with those near the C T D stations and the end-points (grid-points 8, 22, 44 and 58) containing more than 5 profiles. Raw backscatter data is converted to residuals using average fit parameters determined from fits to all profiles in the June, 1993, cruise with bottom depths greater than 320 m. Profiles have been cut at 80% of the minimum of 5 recorded depths as described in Chapter 4.2.4. Absolute currents are determined from bottom-tracking velocities only. Profiles with erroneous ship velocity estimates are flagged, then discarded. Bottom-tracking gave erroneous estimates of ship velocity at the southeastern corner of the survey line (grid-points 3, 5, 6 and 8) for 31 of 591 profiles, with 2 during the 2 n d leg, 8 during the 6th leg and 21 during the 8th and 9th legs. The following sub-sections describe the backscatter and current data observed during the survey period. Chapter 5.1.3 examines the backscatter residuals in plots of all com-binations of the three variables (d, t, z). It closes with an examination of the last hour of the survey period. Chapter 5.1.4 examines the current data as vertical sections in (d, z) to outline the vertical shears, and as horizontal time series in (d, t) to outline the temporal variations. It also discusses the tidal currents predicted and observed during the survey period. Chapter 5.1.5 summarizes the plankton and current structures and their correlations. Chapter 5. Results 81 5.1.3 Backscatter Residuals Quasi-instantaneous vertical sections Figures 5.4a-j are contoured vertical sections in (e?,-, Zfc)-space of the backscatter residuals averaged onto the grid. These sections are not exactly instantaneous, as each spans a time between 1 and 3 hours (except the 10th leg, which only spans 30 minutes). The bathymetry is determined from the D T F only, and so is only a rough approximation to the actual bottom. The 4th leg and the 6th leg have data missing at d ~ 5km, when the ship was closer to the southernmost grid row. These data were not included to ensure little along-shelf variability in backscatter in the vertical sections. The structure during the first 9 hours (Figures 5.4a-d; legs 1-4) remained approxi-mately the same. High plankton concentrations were observed near the ocean surface in a layer 20-30 m thick over the continental slope (southwest sector) and 70-100 m thick over the shelf (northeast sector). A 'depletion layer' filled the area below the surface scattering layer, with scattering levels 15-20 dB lower than at the surface. The extent (in depth) of this depletion layer in the southwest over the slope was from 50 to 100 m. In the center of the line over the shelf break, the depletion layer extended from the surface to 150 m. In the northeast sector of the line over the shelf, the depletion layer extended from 50 m to the bottom. This depletion layer was found throughout the coastwide sur-vey discussed in Chapter 5.2. The deep water seaward of the shelf break contained the highest concentrations of scatterers, reaching levels 10-15 dB higher than at the surface. These deep water scatterers were concentrated primarily at a depth of about 200 m, 4-5 km seaward of the shelf break, in slope water 500-700 m deep. During the first two legs, the near-surface scattering layer in the northeast sector occupied the upper 80 m and extended 4-5 km along the survey line. It moved seaward during the first four legs, covering the surface waters out to the shelf break by about Chapter 5. Results 82 10:08 : ship direction — -> 11:52 o H 0 5 10 15 Distance frow SW corner (km) Figure 5.4: Backscatter residuals (dB) for the June, 1993 A D C P line, (a) Leg 1. Collec-tion times are from 10:08 - 11:52 P D T on 29/06/93 with the ship steaming onshore. Chapter 5. Results 83 Figure 5.4: (b) Leg 2. Collection times are from 11:52 - 14:56 P D T on 29/06/93 with the ship steaming offshore. Chapter 5. Results 84 Figure 5.4: (c) Leg 3. Collection times are from 14:58 -ship steaming onshore. 16:34 P D T on 29/06/93 with the Chapter 5. Results 85 Figure 5.4: (d) Leg 4. Collection times are from 16:36 - 19:14 P D T on 29/06/93 with the ship steaming offshore. Chapter 5. Results 86 19:16 ship direction > 20:54 o H 0 5 10 15 Distance frow SW corner (km) Figure 5.4: (e) Leg 5. Collection times are from 19:16 - 20:54 P D T on 29/06/93 with the ship steaming onshore. Chapter 5. Results 87 Figure 5.4: (f) Leg 6. Collection times are from 20:56 - 23:24 P D T on 29/06/93 with the ship steaming offshore. Sunset occurred at 21:28 PDT, as shown by the dotted line. Chapter 5. Results 88 Figure 5.4: (g) Leg 7. Collection times are from 23:26 P D T on 29/06/93 to 01:06 P D T on 30/06/93 with the ship steaming onshore. Chapter 5. Results 89 03:28 < ship direction 01:06 0 5 10 15 Distance frow SW corner (km) Figure 5.4: (h) Leg 8. Collection times are from 01:06 - 03:28 P D T on 30/06/93 with the ship steaming offshore. Chapter 5. Results 90 5 10 Distance frow SW corner (km) Figure 5.4: (i) Leg 9. Collection times are from 03:28 - 05:20 P D T on 30/06/93 with the ship steaming onshore. Sunrise occurred at 5:24 PDT, just after, the ship reached the end of the section Chapter 5. Results 91 Figure 5.4: (j) Leg 10. Collection times are from 05:20 -occurred at 5:24 P D T , as shown by the dotted line. 05:50 P D T on 30/06/93. Sunrise Chapter 5. Results 92 17:30 during the 4th leg. This surface scattering layer expanded between 12:00 on the second leg and about 16:00 on the third leg to fill the entire water column to a depth of 130 m and over a distance of 6 km. The cross-shelf movement of the shelf scattering layer was related to strong cross-shelf currents, as will be shown in Chapter 5.1.5. The scattering layer over the slope moved up slightly during the fourth leg, from a depth of 200 m to a depth of about 180 m. The backscatter structure during the 5th leg was somewhat similar to that observed during the first four legs. The slight vertical rise in the deep water scatterers during the 4th leg continued during the 5th leg, lifting the centre of the scattering layer to about 165 m depth. A depletion of the shelf water scatterers occurred in the bottom 40 m at a distance of d « 10 km, and in the bottom 100 m further to the northeast (at a distance d=13-14 km). Sunset occurred during the sixth leg of the cruise (Figure 5.4f) and the planktonic structure changed drastically. The upper 50 m contained higher backscatter levels, with the largest concentration above the 250 m isobath at d « 4 km. The intensity of the scattering layer over the slope was reduced by 10-15 dB, and had moved to lie within about 2-3 km of the shelf break. The mid^depth depletion layer observed during the first 5 legs was nearly gone, with only a small area centered at about .150 m depth, at d « 3 km. The structure remained approximately the same for the next 6 hours (legs 7-9; Fig-ures 5.4g-i). High concentrations filled the upper 50 m across the entire survey line. The deep water scatterers remained at a constant level and stationary near the bottom as in the 6th leg. The only major difference was a depletion layer which formed between the sixth and seventh legs of the cruise, extending over 100-150 m depth in the shallow water from the northeast corner out to 2 km beyond the shelf break. This depletion layer was reduced during the 8th leg and had disappeared by the time the ship returned over the Chapter 5. Results 93 shelf during the 9th leg. A strong scattering layer appeared at a depth of about 30 m over the shelf near the northeast corner of the survey line during the ninth leg. This aggregation was only about 2-3 km wide and had backscatter levels about 20 dB higher than typical shallow water levels, which corresponds to a 100-fold increase in acoustic cross-section. It was not present at 01:00-02:00 during the 8th leg and had disappeared by 05:30 when the ship returned on the 10th leg. The 10th leg revealed, instead, a strong scattering layer just above the bottom at a depth of about 110 m. These bottom scatterers in the northeast . corner of the survey line over the shelf appeared first during the 7th leg, with backscatter levels of about + 10 dB, remained at this level during the 8th and 9th legs,.intensifying to about + 25 dB in the early morning hours between the 9th and the 10th legs. The strong scattering layer which appeared in the surface waters during the ninth leg is thought to be a patch of euphausiids which was advected into the survey area between 02:00 and 04:30, and which subsequently descended to the deep waters near the bottom just prior to sunrise (05:30). In Chapters 5.1.4 and 5.1.5, it is shown that the appearance : of the patch was correlated with a reversal of the local water flow from southeastward to northwestward. The composition of the patch will be discussed further in Chapter 5.1.5. V e r t i c a l t i m e series To get a better idea of the major changes occurring in the survey region near sunset, the data were plotted as time series for each grid point, or S'v(zk, tj) for fixed di. Figures 5.5a-e are contoured time series for each of five grid-points 8, 10, 14, 24, and 48, which are located at survey line distances of d=l , 1.5, 2.5, 7.5, and 19.5 km, and over water depths of 600, 500, 400, 250, and 170 m, respectively. Sunset occurred at t = l 1.3 hours. Sunrise occurred at t=19.3 hours, which is only relevant for grid-point 48, since data were only taken in the early morning at the northeastern corner of the survey line (grid-points Chapter 5. Results 94 48-64). Grid :points 8, 10 and 14 over the continental slope revealed nearly identical variability with depth and time. A strong deep scattering layer was present near depths of 200 m from t=0 to t=7 hours, with residuals of 20-30 dB. The upper layers (0-50 m) during this period had scattering levels of about 10 dB, whereas the mid-depths (50-150 m) had levels about 20 dB lower, or about -10 dB. The structure underwent a major change after about 18:00 (t=8 hours). The deep water (>150 m) scattering levels were reduced by about 20 dB, whereas the surface waters underwent a 10-15 dB increase and the mid-depth waters underwent a 20-25 dB increase. The change in structure at t = 8 hours in the time series for Grid-points 8, 10 and 14 is attributed to an upwards migration of zooplankton. The possibility of horizontal advection of zooplankton into the depths between 50 and 150 m can be ruled out. The deep scattering layer, which extended over about 50 m prior to 18:00, rose to occupy the upper 100 m by sunset (t=11.3 hours). This sets the mean vertical swimming speeds at 1-1.5 cm/s. Vertical migrations on these speed and distance scales are common for euphausiids. Hardy and Bainbridge (1954) measured the average swimming speed of a euphausiid at 2.5 cm/s. Pieper (1971) observed euphausiids in Saanich Inlet, British Columbia migrating upwards at sunset with mean speeds of about 2 cm/s. Maximum swimming speeds for euphausiids have been measured by Kampa (1975) at 5-6.5 cm/s. Simard and Mackas (1989) observed Euphausia pacifica on the Vancouver Island conti-nental margin occupying daytime depths of 100-150 m (on the shelf) and 150-200 m (at the shelf break), and rising to occupy the surface waters at dusk. Bollens et al. (1992) observed Euphausia pacifica in Puget Sound occupying 50-125 m during the day and the upper 25 m at night. Genin et al. (1988) observed Euphausia pacifica in the Southern California Bight residing below 200 m during the day and rising to the upper 50 m at night. Chapter 5. Results 95 Figure 5.5: (a) Time series of backscatter residuals S'v(zk, o?;)(dB) at grid-point-8. The water depth is shown with a solid horizontal line at depth = 413 m. Chapter 5. Results 96 Figure 5.5: (b) Same as Figure 5.5a but at grid-point-10, water depth =. 445 m. Chapter 5. Results Figure 5.5: (c) Same as Figure 5.5a but at grid-point-14, water depth = 367 m. Chapter 5. Results Survey Leg 1 2 3 4 5 6 7 8 9 J 1 I I I 1 I I L sunrise i 1 , : — — 1 1 • 1 0 5 10 15 20 Time in hours from 29/06/93 10:10 PDT Figure 5.5: (d) Same as Figure 5.5a but at grid-point-24, water depth = 205 m. Chapter 5. Results Survey Leg 5 6 Time in hours from 29/06/93 10:10 PDT Figure 5.5: (e) Same as Figure 5.5a but at grid-point-48, water depth = 168 m. Chapter 5. Results 100 The vertical migration at sunset is apparent in the time series for grid-point-24 as well (Figure 5.5d). The top of the deep scattering layer is visible for a few hours prior to sunset at about 150 m. This scattering layer is no longer apparent after t=11 hours. The surface water concentrations of scatterers increased by 10-15 dB about an hour prior to sunset, and remained at this higher level for the remainder of the time series. This nighttime shallow water scattering layer extended from the surface down to about 100 m depth. Grid-point-48, over the shelf, had a different scattering layer variability than the profiles over the slope. During the first 5 hours (survey legs 1-3), the concentrations at depths of 80-100 m increased by about 15 dB, as observed in Figure 5.4b and Figure 5.4d, while the scattering levels in the shallower waters remained constant. A slight vertical migration of the deep water scatterers started at about t=7 hours, simultaneously with an increase in surface water scattering levels of about 10 dB. The most remarkable feature of this time series is the strong scattering layer which appeared abruptly at t=16 hours (01:00 on 30/06/93). This feature also appears in Figure 5.4i. It is visible as well in similar time series for grid-points 50, 52, 54 and 56. As discussed above, this patch is thought to have been advected into the survey area and then to have descended into the near-bottom waters just prior to sunrise. This downward migration is visible in. Figure 5.5e, with a strong scattering layer (25 dB) appearing at t=19.5 hours in the water below 130 m. Last hour time series In order to see more clearly what happened during the last hour of the survey period, the profiles taken from 04:36 to 05:50 in the morning of the 30th of June, 1993 are examined more closely. Figure 5.6 is a map of the ship track during this 74 minute period, with the recording times for each profile, and the approximate length scales important for the Chapter 5. Results 101 -125.770 -125.765 -125.760 -125.755 Longitude Figure 5.6: Ship track for the last 74 minutes of the survey period. P D T times are given as well as approximate length scales. analysis. The ship traveled up an ' L ' shaped path from 04:36-05:16, then back down the same path from 05:18-05:40, and returned up nearly the same way from 05:40-05:50. The ship moved over a distance of about 3 km during the first 14 minutes, and again over the last 30 minutes. However, from 04:54-05:20, the ship was stationary to within 500 m. At any given time, horizontal homogeneity is assumed for the region covered by the ship. This assumption allows the backscatter residual profiles to be contoured as a time series (Figure 5.7). The strong scattering layer was located in the upper 70 m during the first 20 minutes of this time series (see also Figures 5.4i and 5.5e). Following twilight start at 04:45, this layer underwent a slight decrease in scattering level (5-10 dB), as well as a slight increase in depth (10-20 m). Just after 05:00, about 20 minutes prior to Chapter 5. Results 102 Figure 5.7: Time series of backscatter residuals (dB) for profiles taken from 4:36 to 5:50 on 30/06/93. Civi l twilight start and upper limb sunrise are shown as dotted lines at 4:45 and 5:24, respectively. Chapter 5. Results 103 sunrise, the scattering layer underwent an abrupt descent, reaching a depth of 100-120 m at about 05:30, or about 10 minutes after sunrise. It remained at this depth near the bottom from 05:30 to the end of the time series at 05:50. The time and length scales of this descent gives mean swimming speeds of 3-4 cm/s. These rates are a factor of two higher than the rates observed for the scattering layer migrating upwards at sunset, and are consistent with Pieper (1971), who observed downwards migratory speeds at sunrise of about 3 cm/s. The major temporal changes in the overall zooplankton structure dominated any spatial variations during this time series, justifying the assumption of horizontal homogeneity. Horizontal time series The final method of observing the various advective and migrational motions of the zooplankton aggregations is to look at horizontal time series, or contour plots of S'v(d,t) for constant z^. Figures 5.8a-i are such horizontal time series, in 24 m depth bins from 6 m to 224 m. The shallow water (6-80 m) scattering levels (Figures 5.8a-c) increased over the south-west half of the survey line (d=2—11 km) by about 10 dB from t=7 to t=12 hours (17:00-22:00). This increase is due to the upwards migration of zooplankton at sunset. The scattering levels over the shelf (from d=11 to d=14 km) remained roughly constant over this time period, but underwent a 5-10 dB increase from t=ll to t=15 hours. A strong scattering layer (also see Figures 5.4i and 5.5e) appeared at d=10-12 km and t=18 hours in the two shallowest bins (6-56 m), disappearing during the 10 16:34 O 5 10 15 Distance frow S W corner (km) Figure 5.11: (c) Leg 3. Collection times are from 14:58 - 16:34 on 29/06/93 with the ship steaming onshore. Chapter 5. Results 127 Figure 5.11: (d) Leg 4. Collection times are from 16:36 - 19:14 on 29/06/93 with the ship steaming offshore. Chapter 5. Results 128 Figure 5.11: (e) Leg 5. Collection times are from 19:16 steaming onshore. - 20:54 on 29/06/93 with the ship Chapter 5. Results 23:24 ship direction -10 •.. ••—Kr ; -eo -20 - . T I O . - - . . . . - ..\"-2d; -10 \\ \\ -•tp '*<<>-.-;.':':':\"::: :-2o- •': -10—. -1.0 \" 7 \" -10 '; I s 5 10 Distance frow S W corner (km) 15 9 - 1 0 . -to -10 •.'-20 5 10 Distance frow S W corner (km) Figure 5.11: (f) Leg 6. Collection times are from 20:56 - 23:24 on 29/06/93 with the steaming offshore. Chapter 5. Results 130 23:26 ship direction : > 01:06 O S 10 15 Distance frow S W corner (km) o O 5 10 15 Distance frow S W corner (km) Figure 5.11: (g) Leg 7. Collection times are from 23:26 on 29/06/93 to 01:06 on 30/06/93 with the ship steaming onshore. Chapter 5. Results 131 0 3 : 2 8 < s h i p d i r e c t i o n 0 1 : 0 8 o 0 5 10 15 D i s t a n c e f row S W c o r n e r (km) o O 5 lO 15 D i s t a n c e f row S W c o r n e r (km) Figure 5.11: (h) Leg 8. Collection times are from 01:06 - 03:28 on 30/06/93 with the ship steaming offshore. Chapter 5. Results 132 o O 5 10 15 D i s t a n c e f row S W c o r n e r (km) Figure 5.11: (i) Leg 9. Collection times are from 03:28 - 05:20 on 30/06/93 with the ship steaming onshore. Chapter 5. Results 133 o 0 5 10 15 Distance frow S W corner (km) Figure 5.11: (j) Leg 10. Collection times are from 05:20 - 05:50 on 30/06/93. Chapter 5. Results 134 during leg 6 also suggest semidiurnal baroclinic tidal motions as were observed in the tangential component. The ninth leg revealed baroclinic shear in the normal component, with weak southeastward flows (w 5-10 cm/s) below 100 m, and weak northwestward flows ( « 0-5 cm/s) in the surface waters (30-100 m). The 10th leg currents were weak and to the northwest throughout the water column, with small patches of water flowing weakly in the opposite direction. S u m m a r y of Currents The overall current pattern was a strong (PS 20 cm/s) southeastward flow throughout the water column during the first 16 hours (legs 1-8). The speeds during the next three hours (legs 9-10) in the shallow waters (0-80 m) over the shelf and shelf break were reversed to weak northwestward flows. The deep water (> 80 m depth) during legs 9-10 still revealed southeastward flows. Vertical shears in tangential component were observed at the northeast corner of the survey line over the shelf during the 1st and 2nd legs, and again in the same area during the 6th and 1th legs. Vertical shears in the normal component were observed over the shelf break and slope during the 1st leg and again during the 6th leg. The reversal in flow from southeastward to northwestward agrees with the predic-tions of the barotropic tidal model (Chapter 5.1.4). However, strong vertical shears over the shelf and shelf break suggest that the tidal motions have strong Is* and 2 n d mode baroclinic components. The current vectors taken in the same area at the same time in Figures 5.10a-e exhibit variability of about ± 5 cm/s in speed and ±20° in direction. These fluctuations were most probably due to the combined effects of random errors in the water and bottom velocity estimates. In Chapter 4.4.3, a rough prediction of these errors yielded au — ov = 5 cm/s. The errors can be calculated from the data taken during the survey line by assuming Chapter 5. Results 135 horizontal homogeneity in the current components over the length scales covered by the ship between profiles (about 500 m). If this assumption holds, then the mean difference in absolute water velocity components between neighboring profiles ((Au) and (Ai>).over the A D C P sampling interval, At) should be zero: '(Aw) = (u(t) - u(t - At)) = 0, (5.1) (Av) = (v(t) - v(t - At)) - 0. (5.2) Therefore, the variances in (Au) and (Av) give estimates of ou and 80 m) for the first 2 legs was correlated with onshore flowing water below 80 m Chapter 5. Results 138 depth. The subsequent reversal to offshore flow sometime after 13:00 or 14:00 en-abled advection of the deep water (> 80 m) scatterers into the survey area along with the surface water (< 80 m) scatterers. 2. During the 6th leg of the survey, a northwestward current developed just over the shelf-slope at about 100 m from the bottom (Figure 5.11f). It extended from d=2-5 km and from a depth of 180 m to 300 m. At the same time and location, an aggregation of zooplankton was observed with the same spatial structure in z and d as the water velocities (Figure 5.4f). This aggregation was originally thought to be a remnant of the deep scattering layer which has already begun its upwards migration. However, the possibility also exists of horizontal advection by localized currents. The near-bottom plankton aggregation remained in its initial position for the remainder of the survey period, whereas the modified current pattern was much weaker in the seventh leg, and completely gone by the eighth leg. Whether the zooplankton were advected in from the south or whether they remained at depth after sunset, they stayed near the bottom just past the shelf break throughout the night, regardless of the current variations. 3. During the 9th leg of the survey, a strong scattering layer appeared in the area, as discussed above (Figure 5.4i). This layer was not present at about 2:00 in the morning during the eighth leg, and migrated quickly downwards to the bottom layers just before sunrise, where it was found during the tenth leg. Therefore, this scattering layer must have been advected into the survey area between 2:00 and 4:00 on the morning of 30/06/93. If it was present at 2:00 just outside the range of the grid-points, it would have been advected by approximately 1 km in the two following hours before the ship returned, giving an estimate for the advection speed of about 10 cm/s. This is roughly corroborated by the current data, which Chapter 5. Results 139 shows speeds normal to the section of about 5-10 cm/s throughout the ninth leg (Figure 5.Hi). A better visualization of the correlation between the appearance of this aggregation of scatterers and the current structure is obtained by plotting the normal currents as a function of time for grid-point-48 (Figure 5.13), and comparing this to Figure 5.5e. The change in normal currents in the depth range 30-80 m from southwestward to northeastward flows is clearly seen at t « 17 hours, and is correlated with an almost identical structure in the zooplankton concentrations in Figure 5.5e. The advection of this high concentration seems certain, leaving only the question of its origin. Advection from the south-east seems to indicate an origin over the small canyon located a few kilometres to the south of the survey line (see Figure 5.2). 5.2 Semi-Synoptic Continental Margin Survey This section examines the currents and zooplankton distributions observed during the June, 1993 cruise within the area defined by the grid in Figure 4.16. The spatial and temporal scales are larger here than for the A D C P survey line described in the last section, and the coverage of data for each location is more sparse. After a brief description of the June 1993 cruise, the data are examined as cross-shelf in-situ vertical sections about 100 km long at various times in June, 1993. The persistent features in these vertical sections are discussed. Then, the data are horizontally averaged to produce maps of the entire southwest Vancouver Island continental margin. Such average distributions must be approached with caution, as high temporal variabilities on diurnal time scales in backscatter residuals due to advection and migration and in water velocities due to tidal flows were observed during the A D C P survey line (Chapter 5.1). A single cruise, which Chapter 5. Results 140 Survey Leg 1 2 3 4 5 6 7 8 9 J i i i i i i i t_ sunset sunrise 0 5 10 15 20 Time in hours from 29/06/93 10:10 PDT Figure 5.13: Grid-point 48 time series of normal currents v(zk,tj) (cm/s). Negative values are contoured with dotted lines, positive values with solid lines. Positive normal components point 'into' the page, in a northwestward direction. Positive tangential components point to the right, in a northeastward direction. Chapter 5. Results 141 typically results in 2-3 sampling periods 1 averaged onto each grid-point, will not contain enough sampling periods to yield accurate seasonal mean values. Thus, the averages can only be used to represent the persistent features found in the in-situ vertical sections over larger horizontal spatial scales. 5.2.1 June, 1993 Cruise and Data During the June, 1993 La Perouse cruise, the C.S.S. J.P. Tully spent 356 hours in the area defined by the grid in Figure 4.16, recording a total of 10680 A D C P profiles. Figure 5.14 shows the ship's tracks within the area defined by the grid in Figure 4.16. Each of eight survey lines perpendicular to the coast are numbered (1-8) in Figure 5.14, and will be discussed later in this section. During the cruise, the ship was collecting water property data and servicing moorings along the west coast from Juan de Fuca Strait to Brooks Peninsula. It entered the area from the mouth of Juan de Fuca Strait at 01:19 on 19/06/93, and proceeded up the coast over the deeper waters, steaming back and forth in lines parallel to the grid rows, covering the first 6-20 columns of the grid on each pass. It left the area from the northwest corner (49°N, 125.0°W) at 03:28 on 22/06/93, proceeding north to Brooks Peninsula. It returned to the northwest corner of the area at 02:14 on 26/06/93 and proceeded to Barkley Sound (48.7°iV, 125.0°W), where it arrived at about 10:00 on 27/06/93. It subsequently steamed along nearly three complete grid rows, before arriving at the A D C P survey line (described in Chapter 5.1) at 10:10 on 29/06/93. Following the survey, it proceeded to the Barkley Sound area again, arriving at 12:00 on 30/06/93. It left at 21:30 on the same day and steamed along one complete grid row on its way to Endeavour Ridge, (46.0°N, 130.0°W), leaving the area at 23:00 on 30/06/93. It returned to the area at 18:00 on 08/07/93 over the slope near the southwest lA 'sampling period' denotes a period of time during which all ADCP profiles were sequential and averaged onto the same grid point Chapter 5. Results 142 —I 1 1 1 1 1 r— -127.5 -127.0 -126.5 -126.0 -125.5 -125.0 -124.5 Longitude -127.5 -127.0 -126.5 -126.0 -125.5 -125.0 -124.5 Longitude Figure 5.14: June 1993 ship tracks within the area of interest. Eight survey lines are numbered (1-8) (see Table 5.1). Arrow heads are spaced at 2 hour intervals along the ship tracks. The area covered by the grid is delineated by a dotted line, (a) From 01:19 on 19/06/93 to 03:30 on 22/06/93, (b) from 02:00 on 26/06/93 to 10:00 on 29/06/93. Chapter 5. Results 143 -127.5 -127.0 -126.5 -126.0 -125.5 -125.0 -124.5 Longitude Figure 5.14: (c) From 10:00 on 29/06/93 to 23:00 on 30/06/93 and from 18:00 on 08/07/93 to 01:00 on 08/07/93. Chapter 5. Results 144 Section Start-End Grid Start-End A * Day or Number Grid Column Row(s) Date;Time (PDT) (hours) Night 1 2-34 10 27/06;ll:44 - 28/06;03:44 12 Both 2 5-34 9 30/06;17:44 - 30/06;22:44 5 Both 3 : 1-21 10-11 19/06;22:28 - 20/06;05:28 7 Night 4 1-21 10-11 20/06;05:28 - 20/06;ll:28 6 Day 5 2-34 6 28/06;07:44 - 28/06;22:44 15 Day 6 8-33 2 29/06;00:44 - 29/06;08:44 8 Both 7 4-36 3-4 19/06;00:27 - 19/06;10:28 11 Both 8 1-35 1-6 07/07;17:58 - 07/07;23:58 7 Both Table 5.1: Eight June vertical sections taken along grid rows and their time and length scales. The ship tracks for these sections are shown in Figure 5.14. corner of the grid, and went directly to the mouth of Juan de Fuca Strait, leaving the area from the south east corner at 01:00 on the same day. The data set described above includes eight vertical sections (numbered 1-8 in Fig-ure 5.14) taken as the ship steamed onshore or offshore. During each one of these sections, the ship usually stayed within about 4 km of a grid row, and covered most of the length of the grid, or about 100 km, in 5 to 15 hours. Table 5.1 presents the temporal and spatial coverage of these eight sections, as well as their alongshore locations iri terms of grid rows, and whether the data was taken at night or during the day. Sections 1 through 4 were taken perpendicular to the coast off the mouth of Barkley Sound. They covered most of the grid length, out to a distance of 100 km from the coast, but only covered about three grid rows, or about 20 km in an alongshore direction. Sections 5-8 covered a larger area towards the south of the grid, spanning the grid length (100 km) and about six grid rows or 50 km in an alongshore direction. The A D C P was set to record 120-second averages of profiles, using a 4 m bin length, a 4 m blank beyond-transmit, and a total of 128 depth bins. Ship speeds during the cruise varied from about 0 knots during water property measurements to about 15 knots as the ship steamed from Barkley Sound to Endeavour Ridge. The spatial coverage of the Chapter 5. Results 145 profiles therefore had a resolution of about 0-1 km, with an average of about 500 m at normal cruising speeds (10 knots). 5.2.2 In-Situ Scattering Layer Distributions Figures 5.15-5.18 are contoured vertical sections of backscatter residuals for sections 1-4, respectively. The horizontal axes at the bottom give the relative distance in kilometres measured from the most westward point in the section. The axes at the top give the consecutive numbers of the grid-points, which are numbered from south to north, west to east. The tick marks on the right axes show the location of 8 m bins. The bathymetry is estimated from Digital Terrain File (DTF) depths only, and so is only an approximation of the true bottom depths. A l l four sections had a similar structure, with the highest scattering levels occurring in the upper 50 m over the entire section length, reaching +25 dB at night and about +15 dB during the day. A depletion region occupied the mid-depths (50-150 m) over the slope, with residual levels of -5 dB to -10 dB. The scattering levels in deep water (150-400 m) were higher than in the mid-depths, usually reaching 5-8 dB. A D C P measurements taken over the shelf break at mid-depth range of 50-150 m along cross-sections 1, 2 and 3 (measured at night) revealed scattering levels comparable to the levels in the surface waters, leading to near-uniform scattering levels throughout the water column over the shelf break. Cross-section 4 (measured during the day), however, revealed depleted scattering levels during the day of -5 dB to -3 dB in the mid-depth range (50-150 m) over the shelf break and a strong scattering layer in the deeper water (150-200 m) over the shelf break. This strong shelf-break bottom scattering layer was probably euphausiids at their daytime depth. Such a scattering layer was not observed in the sections where the ship steamed over the shelf break at night (sections 1 and 3 - Figures 5.15 and 5.17), when the euphausiids had presumably risen to the Chapter 5. Results 146 •« • ship direction ss night j day PROFILE LABELS & LOCATIONS ojl \\ & .-71*1 ^ ° b<& tfl tf£ RELATIVE DISTANCE (km) (West to East) Figure 5.15: Contoured backscatter residuals (dB) for section 1. Data was collected from 11:44 P D T on 27/06/93 to 03:44 P D T on 28/06/93, with the ship steaming offshore. Sunset occurred at 21:30 P D T on 27/06/93, when the ship was over the shelf break, at a relative distance of « 40km, as indicated by the dashed arrow. Chapter 5. Results 147 o o m I o o o co <— 3 1 C N X I U J T) Q C N o o O m I o o I night ss 0 — ship direction : day P R O F I L E L A B E L S &c L O C A T I O N S A 4 , 0 ^ < ^ 6 ^ ^ bfC1 ^ 3 0 4 5 6 0 7 4 8 9 R E L A T I V E D I S T A N C E ( k m ) ( W e s t t o E a s t ) 1 0 4 co U J I < 0 0 o O < O o Figure 5.16: Contoured backscatter residuals (dB) for section 2. Data was collected from 17:44 to 22:44 P D T on 30/06/93, with the ship steaming offshore. Sunset occurred at 21:30 P D T , when the ship was over the deep water, at a relative distance of w 15fcm, as indicated by the dashed arrow. Chapter 5. Results 148 o I o o o cn _ fe I CN Q - 9 U J U O I o o to o LO o o I \\ 0 qfi — ship direction — — night PROFILE LABELS & LOCATIONS sr CO UJ < CO o o I— < o o 10 2 0 29 39 49 59 RELATIVE DISTANCE (km) (West to Eas t ) • Figure 5.17: Contoured backscatter residuals (dB) for section 3. Data was collected from 22:28 P D T on 19/06/93 to 05:28 P D T on 20/06/93, with the ship steaming onshore. Sunrise occurred at 05:24 P D T on 20/06/93, just as the ship reached the end of the section, as indicated by the dashed arrow. Chapter 5. Results 149 Figure 5.18: Contoured backscatter residuals (dB) for section 4. Data was collected from 05:28 to 11:28 P D T on 20/06/93, with the ship steaming offshore. Sunrise occurred at 05:24 P D T on 20/06/93, just before the ship started the section, as indicated by the dashed arrow. Chapter 5. Results 150 surface. A shelf-break bottom scattering layer was not observed in cross-section 2 either (Figure 5.16), even though the ship steamed over the shelf break during the day. However, the ship was moving at about 20 knots along section 2, and so it steamed over the shelf-break about an hour prior to sunset. The euphausiids over the shelf-break had probably started their upwards migration, accounting for the lack of a shelf-break bottom scattering layer at this time. The water over the shelf in sections 1 and 2 had a uniform structure, with scattering levels for the entire water column ranging between 8 and 15 dB. A slight depletion region (levels 3-5 dB lower than in the surrounding water) was observed in both sections over the 100 m deep basin at relative distances of 70 to 90 km. Figures 5.19-5.22 are contoured vertical sections of backscatter residuals for sections 5-8, respectively. The parts of these sections over the shelf break and slope were taken during daylight, and had a similar structure as in section 4 (see Figure 5.18). Aggre-gations of scatterers were found at the surface and at depths of 150-250 m near the shelf-break. A depletion layer occupied the depths from 50-150 m over the shelf-break and slope, and extended a few kilometres onto the shelf. The resulting definite sepa-ration between the shallow-water and the deep-water scattering layers was most likely due to euphausiids at their daytime depths over the shelf break, as observed in section 4. The structure over the shelf in sections 5-8 was again uniform, as in sections 1 and 2 (Figures 5.15 and 5.16). The only high (> 15dB) scattering layer was in section 5 close to the bottom over a shallow basin inthe shelf, at. relative distances of 70-90 km. 5.2.3 In-Situ Current Patterns The errors in the absolute water velocities due to inaccuracies in ship velocity estimates can be calculated in a manner similar to that in Chapter 5.1.4. The currents are assumed uniform (to within 1 cm/s) over the length scales covered by the ship in an averaging pe-riod, so the differences in current components between neighboring profiles give variances Chapter 5. Results 151 Figure 5.19: Contoured backscatter residuals (dB) for section 5. Data was collected from 07:44 to 22:44 P D T on 28/06/93, with the ship steaming onshore. Sunset occurred at 21:30 P D T on 28/06/93 when the ship was over the shelf at a relative distance of « 105 km, as indicated by the dashed arrow. Chapter 5. Results 152 o L O I o o o ^ L O CO , — 2 ' C M x I U J m Q C M o o rO I o L O I o o I day [ night PROFILE LABELS & LOCATIONS ship direction <5-2 3 3 5 47 5 8 7 0 RELATIVE DISTANCE (km) (West to Eas t ) Figure 5.20: Contoured backscatter residuals (dB) for section 6. Data was collected from 00:44 to 08:44 P D T on 29/06/93, with the ship steaming offshore. Sunrise occurred at 05:30 P D T when the ship was over the shelf at a relative distance of « 25km, as indicated by the dashed arrow. Chapter 5. Results 153 ship direction sr day j night PROFILE LABELS & LOCATIONS RELATIVE DISTANCE (km) (West to Eas t ) Figure 5.21: Contoured backscatter residuals (dB) for section 7. Data was collected from 00:27 to 10:28 P D T on 19/06/93, with the ship steaming offshore. Sunrise occurred at 05:30 P D T when the ship was over the shelf at a relative distance of 40km, as indicated by the dashed arrow. Chapter 5. Results 154 o LO I o o o CO v— V I | o CM x I £^ uj m I o o I o m f~> I o o I 1 V • W — ship direction ss day | night PROFILE LABEJLS & LOCATIONS 6,6^ \\ ^ r oo-CL < 00 o o I— e ^ 0 * &0^> lofe\"^ - i V i -es ' O 1 1 2 2 3 3 4-4 5 5 6 5 7 6 R E L A T I V E D I S T A N C E ( k m ) ( W e s t to E a s t ) Figure 5.23: Contoured normal (top) and tangential (bottom) currents (cm/s) for sec-tion 1. Normal currents are positive to the northwest, tangential currents are positive to the northeast. Data was collected from 11:44 P D T on 27/06/93 to 03:44 P D T on 28/06/93, with the ship steaming offshore. Chapter 5. Results 158 P R O F I L E L A B E L S Sc L O C A T I O N S O 11 21 3 2 4-3 54- 64- 7 5 R E L A T I V E D I S T A N C E ( k m ) (Wes t to E a s t ) P R O F I L E L A B E L S & L O C A T I O N S O 1 1 2 1 3 2 4-3 54- 64- 7 5 R E L A T I V E D I S T A N C E ( k m ) (Wes t to E a s t ) Figure 5.24: Contoured normal (top) and tangential (bottom) currents (cm/s) for sec-tion 2. Normal currents are positive to the northwest, tangential currents are positive to the northeast. Data was collected from 17:44 to 22:44 PDT on 30/06/93, with the ship steaming offshore. Chapter 5. Results 159 P R O F I L E L A B E L S 8c L O C A T I O N S a - i O a < b 7 . ^gfc ^ -t^^O ^zfL ^Qlo ^ > 7-O 5 9 14- 19 2 3 2 8 3 3 R E L A T I V E D I S T A N C E ( k m ) ( W e s t to E a s t ) P R O F I L E L A B E L S 8c L O C A T I O N S rgP- TTI * a \" 3 ^ « T l ^ O 5 9 14- 19 2 3 2 8 3 3 R E L A T I V E D I S T A N C E ( k m ) (Wes t to E a s t ) Figure 5.26: Contoured normal (top) and tangential (bottom) currents (cm/s) for sec-tion 4. Normal currents are positive to the northwest, tangential currents are positive to the northeast. Data was collected from 05:28 to 11:28 P D T on 20/06/93, with the ship steaming offshore. Chapter 5. Results 161 P R O F I L E L A B E L S Sc L O C A T I O N S R E L A T I V E D I S T A N C E . ( k m ) (Wes t to E a s t ) P R O F I L E L A B E L S &c L O C A T I O N S ' O T l 2 1 3 2 4-3 54- 6 4 R E L A T I V E D I S T A N C E ( k m ) ( W e s t to E a s t ) Figure 5.27: Contoured normal (top) and tangential (bottom) currents (cm/s) for sec-tion 5. Normal currents are positive to the northwest, tangential currents are positive to the northeast. Data was collected from 05:28 to 11:28 P D T on 20/06/93, with the ship steaming offshore. Chapter 5. Results 162 P R O F I L E L A B E L S Sc L O C A T I O N S r£&- qjcfr -lp<° 6,7-0 ^ ~ feSA O 11 21 3 2 4 3 5 4 6 4 7 5 R E L A T I V E D I S T A N C E ( k m ) (West to E a s t ) P R O F I L E L A B E L S Sc L O C A T I O N S • O 11 21 3 2 4 3 5 4 6 4 7 5 R E L A T I V E D I S T A N C E ( k m ) (West to E a s t ) Figure 5.28: Contoured normal (top) and tangential (bottom) currents (cm/s) for sec-tion 6. Normal currents are positive to the northwest, tangential currents are positive to the northeast. Data was collected from 05:28 to 11:28 PDT on 20/06/93, with the ship steaming offshore. Chapter 5. Results 163 P R O F I L E L A B E L S Sc. L O C A T I O N S O 12 2 5 3 7 4-9 61 74- 8 6 R E L A T I V E D I S T A N C E ( k m ) ( W e s t to E a s t ) P R O F I L E L A B E L S & L O C A T I O N S O 12 2 5 3 7 4 9 61 7 4 8 6 R E L A T I V E D I S T A N C E ( k m ) ( W e s t to E a s t ) Figure 5.29: Contoured normal (top) and tangential (bottom) currents (cm/s) for sec-tion 7. Normal currents are positive to the northwest, tangential currents are positive to the northeast. Data was collected from 05:28 to 11:28 P D T on 20/06/93, with the ship steaming offshore. Chapter 5. Results 164 P R O F I L E L A B E L S Sc L O C A T I O N S O l O 21 31 4-1 51 6 2 7 2 R E L A T I V E D I S T A N C E ( k m ) ( W e s t to E a s t ) P R O F I L E L A B E L S 8c L O C A T I O N S O 10 2 1 3 1 4-1 51 6 2 7 2 R E L A T I V E D I S T A N C E ( k m ) ( W e s t to E a s t ) Figure 5.30: Contoured normal (top) and tangential (bottom) currents (cm/s) for sec-tion 8. Normal currents are positive to the northwest, tangential currents are positive to the northeast. Data was collected from 05:28 to 11:28 P D T on 20/06/93, with the ship steaming offshore. Chapter 5. Results 165 Section 8 revealed a shear in the normal component over Juan de Fuca Canyon at relative distances of 65-75 km of about 50 cm/s over 200 m, with normal flows reversing from northwestwards in the upper 50 m to southeastward in the deeper waters. 5.2.4 Spatially Averaged Zooplankton Distributions The zooplankton distributions over the entire southwest Vancouver Island continental margin can be examined by averaging the June 1993 cruise onto the 22 x 38 grid shown in Figure 4.16, and then into 24 m depth bins. The data initialized 326 of 836 grid-points (39%), with vertical coverage extending from 6 metres depth to about 80% of the bottom depth. Such an average combines data from different times of the month of June, 1993, and therefore must be approached with caution. Large variations in zooplankton distributions on diurnal time scales were observed in Chapter 5.1, especially at times near sunrise and sunset and at locations near the shelf break. Such variability makes an estimate of the seasonal mean zooplankton distribution very difficult, since each grid-point's averaged scattering levels will include only one or two different sampling periods. Typically, at least 3-5 sampling periods are needed to resolve diurnal variations. Thus, in the following analysis, only those large scale patterns observed to be persistent in Chapter 5.2.2 are examined. Any smaller scale variability is not reliable with so little data in the average. Figures 5.31, 5.33 and 5.35 are horizontal maps of the daytime backscatter residuals averaged over the depth ranges [6,30] m, [78,102] m and [174,198] m, respectively. Fig-ures 5.32, 5.34 and 5.36 are maps of the averaged nighttime backscatter residuals for the same depth ranges. Note that \"absolute\" backscatter residual levels cannot be compared between daytime and nighttime as different average fit parameters are used for day and night residual calculations. The upper 30 m revealed a horizontal structure which was roughly uniform (to within Chapter 5. Results 166 127 3 0 • 126 • 3 0 . 125 Figure 5.31: Vertically-averaged (6-30 m) daytime backscatter residuals (dB) for the period June 19 - July 7, 1993. Chapter 5. Results 167 127 3 0 126 . 3 0 . 1 2 5 Figure 5.32: Vertically-averaged (6-30 m) nighttime backscatter residuals (dB) for the period June 19 - July 7, 1993. Chapter 5. Results 168 127 . 3 0 126 3 0 125 Figure 5.33: Vertically-averaged (78-102 m) daytime backscatter residuals (dB) for the period June 19 - July 7, 1993. Chapter 5. Results 169 127 3 0 126 3 0 . 1 2 5 Figure 5.34: Vertically-averaged (78-102 m) nighttime backscatter residuals (dB) for the period June 19 - July 7, 1993. Chapter 5. Results 170 127 3 0 126 3 0 125 Figure 5.35: Vertically-averaged (174-198 m) daytime backscatter residuals (dB) for the period June 19 - July 7, 1993. Chapter 5. Results 171 127 3 0 126 \\ 3 0 125 Figure 5.36: Vertically-averaged (174-198 m) nighttime backscatter residuals (dB) for the period June 19 - July 7, 1993. Chapter 5. Results 172 ±10 dB) both night and day, with backscatter residuals of about 20 dB during the day, and about 10-15 dB during the night. A slight depletion region along the shelf break was apparent both day and night in the surface waters, with levels « 5dB lower than the levels over the shelf and over the slope. The mid-depth water (78-102 m) revealed a more banded structure along the shelf-break both day and night, with higher scattering levels over the shelf (15-20 dB during the day and 5-10 dB at night) than over the shelf-break and slope (—10-0 dB during the day and —5-0 dB at night). Lower scattering levels in the water from 78-102 m correspond to the depletion layer at the same depths seen in most vertical sections examined in the previous sections (e.g. Figures 5.4a-e and 5.15-5.22). The deep water (174-198 m) also revealed a banded structure along the shelf break. Scattering levels over the shelf break at night were 5-10 dB lower than the surface water levels. In contrast, the scattering levels over the shelf break during the day were about 15 dB lower than the surface water levels, due to plankton rising to their daytime depths. Figure 5.37 shows two schematics of the cross-shelf zooplankton structure present along most of the shelf break in June, 1993 (a) at night and (b) during the day. These schematics show the general features observed in zooplankton distributions in Chapter 5.1 and in Chapter 5.2. High concentrations of zooplankton resided in the surface waters and a depletion region filled the depths from 100-200 m both day and night. Strong con-centrations of scatterers resided just over the shelf break at depths of 150-200 m during the day. This shelf-break bottom scattering layer underwent migration into the surface waters at sunset, descending back to its\" daytime depth at sunrise. The surface scatterers are thought to be comprised of mainly copepods during the day, and of copepods and euphausiids at night. The daytime bottom shelf-break scattering layer is thought to be comprised of mainly euphausiids. Chapter 5. Results 173 Figure 5.37: Schematic representations of the zooplankton structures found along the shelf-break in June, 1993 (a) at night. Dark areas indicate high backscatter residuals. Chapter 5. Results 174 Chapter 5. Results 175 5.2.5 Spatially Averaged Current Patterns The current patterns over the entire southwest Vancouver Island continental margin are also examined by averaging the June 1993 cruise onto the 22 x 36 grid shown in Figure 4.16, and then into 24 m depth bins. Such an average combines data from different times of the month of June, 1993, and therefore must be approached with caution. Large variations in water flows on diurnal and semi-diurnal time scales due to strong tidal currents were observed in Chapter 5.1. Such variability makes an estimate of the seasonal mean flows very difficult, since each grid-point's averaged water velocities will include only one or two different sampling periods. Typically, about 6-10 different sampling periods would be needed in each average to resolve semi-diurnal variations (M.Foreman, pers. comm.). Thus, in the following analysis, only those large scale patterns observed to be persistent in Chapter 5.2.3 are examined. Any smaller scale variability is not reliable with so little data in the average. Figures 5.38 and 5.39 are horizontal maps of the absolute current vectors averaged over the depth ranges [6,30] m and [78,102] m, respectively. Absolute currents from GPS estimates of ship velocity are included in these maps. Two predominant features were observed. First, a strong flow (w 20 cm/s) to the south-southwest over the shelf break (200 m isobath), as was observed in the vertical sections examined above. This is consistent with observations of a southwestward summer shelf break current by Freeland et al. (1984). Second, a strong outflow (> 30 cm/s) to the northwest from Juan de Fuca Strait, which appeared to be turning to the south and southeast over Juan de Fuca Canyon. This outflow is apparent in the upper 30 m, but is much reduced at 80 m. This is consistent with the two-layer estuarine flow in the strait described by Labrecque et al. (1994), who found a seaward surface flow of relatively fresh water in the upper 80 m, and a landward flow of more saline oceanic Chapter 5. Results 176 127 3 0 126 3 0 125 Figure 5.38: Vertically-averaged (6-30 m) horizontal current vectors for the period June 19 - July 7, 1993. Chapter 5. Results 177 127 3 0 126 3 0 125 Figure 5.39: Vertically-averaged (78-102 m) horizontal current vectors for the period June 19 - July 7, 1993. Chapter 5. Results 178 water at depth. A small clockwise eddy also appeared just north of the outflow from Juan de Fuca Strait. The flows over the slope are weak and confused, as expected from the errors in the GPS estimates of ship velocity. The flows over the shelf are also weak and confused, due primarily to a lack of temporal coverage and tidal velocities which are of the same magnitude as the mean velocities. Some evidence of a northwestward flowing coastal current is apparent near the coast off Cox Point (Tofino) at (49°iV,125.7°VK). A schematic of the major surface water flow patterns observed in Figure 5.38 is presented in Figure 5.40. Chapter 5. Results Chapter 6 Summary and Recommendations Goals The goals of this work were first, to develop a consistent protocol for estimating backscatter strengths and current velocity from vessel-mounted A D C P data; and second, to use this protocol to examine spatial and temporal variability of zooplankton distribu-tions off the southwest coast of Vancouver Island. Each of these goals will be addressed in the following summary, including what limitations the A D C P data have in making estimates of the backscatter strengths and currents, and what changes could be made to the present system to improve the data quality and resolution. Estimating absolute backscatter values Absolute measures of backscatter were difficult to obtain due to variability in the system parameters used to convert the raw A G C to backscatter strength, Sv. The pa-rameters with the largest effects were the electronics noise, Er, the power input to the transducer, Vs, and the temperature of the electronics, Te. These parameters introduced errors in the estimates of Sv of about 5 dB. A measurement of Er and of V^, and continu-ous monitoring of Te should reduce the error on Sv to below 1 dB. Using such an estimate of Sv, it would be possible to calculate zooplankton biomass using a model of how each species scatters 150 kHz sound and measurements of the relative abundances and weights of zooplankton from net samples taken coincidentally with the A D C P measurements. 180 Chapter 6. Summary and Recommendations 181 Residuals Most of the system-dependent effects in the backscatter profiles can be removed if the data are only used as a relative measure of zooplankton biomass. This is achieved with the residuals obtained after linear regression analysis on the data. Average regression parameters from 10000 profiles remove all system dependent effects from the backscatter residuals to within about 2 dB. This estimate of the errors in backscatter residuals de-pends on maximal variations in the system parameters affecting the backscatter strength (Table 4.2), and is probably smaller than 2 dB. A continuous record of the system pa-rameters in Table 4.2 would yield a more precise estimate of the errors in backscatter. residuals. Bottom-scattering effects Because averaging intervals often covered substantial variations in bottom depths, sidelobe bottom scattering effects in the vessel-mounted A D C P used in this study of-ten contaminated a greater fraction of the water column than the specified 15% (RDI, 1989). Bottom-scattering effects by sidelobes were removed by using the minimum of five estimates of local water depth: Three recorded by the A D C P bottom-tracking pings; and two independent estimates from the Digital Terrain File. Such a method sometimes removes more of the profile than is necessary. However, this is a sacrifice which must be made in order to avoid ambiguities in the resulting backscatter profiles near the bottom. More accurate removal of bottom scattering effects could have been achieved using one of the following methods: 1. Record the minimum of all the depths measured by the A D C P bottom tracking pings in each ensemble averaging interval. In the present configuration, only the average bottom depth of the 20-30 depths was recorded. A simple algorithm could be implemented to record the absolute minimum depth encountered during each Chapter 6. Summary and Recommendations 182 averaging interval. 2. Implement a sonar beam that points vertically downwards in the A D C P (a 5-beam instrument). The bottom profile measured with a vertical beam is not affected by side-lobe bottom scattering. The vertical beam could therefore measure accurate backscatter strengths to within one bin length of the bottom. It could also be used in concert with the other four beams to measure backscatter strengths in the upper 80% of the water column. Gargett (1988) used a single narrow-beam 215 kHz transducer mounted on the vessel pointing straight down. Such a device could be implemented on the C.S.S. J.P.Tully and used with the vessel-mounted A D C P already in place to measure scattering strengths close to the bottom. Separating profiles into day/night It is important to distinguish between profiles taken during the day and during the night, as the zooplankton distributions change drastically at sunset and at sunrise. Upper limb sunset and sunrise times obtained from astronomical data were chosen as delimiters between day and night. Upper limb sunset and sunrise times are given in U T C , and so it is important to know what time zone the times recorded by the A D C P refer to. Unfortunately, the time zone is not recorded by the data acquisition system (DAS), and so must be found by examining the scientist's log-book in concert with the A D C P data. A record of the time zone referred to by the A D C P would be helpful. Errors in water velocities The errors in the absolute water velocities were estimated using the RDI (1989) method to be about 5 cm/s when using bottom-tracking estimates of ship velocity. The errors in absolute water velocities from GPS estimates of ship velocity were estimated to be about 40 cm/s. The bottom-tracking errors were small enough to yield accurate Chapter 6. Summary and Recommendations 183 measurements of local currents, especially when spatially averaged over distances of 4-8 km (which reduces the errors to about 2 cm/s). The GPS errors, on the other hand, were about twice the typical mean currents for the southwest coast of Vancouver Island, and as such were not very useful. Spatial averaging over distances of 4-8 km reduced the errors to about 15-20 cm/s, which is of the same order of magnitude as the local currents. Broad spatial averages over distances of about 30 km would be necessary to reduce the errors in absolute currents from GPS estimates of ship velocity to acceptable levels (5 cm/s). The use of differential GPS or removal of the selectivity mode by the U.S. Navy would improve positional accuracy and hence velocity by about an order of magnitude. The error estimates in absolute water velocities given in the preceding paragraph must be regarded with some skepticism as they do not include the systematic errors due to the the offset angle and scaling bias of the A D C P mounted on the J.P. Tully. These effects have never been measured and were not corrected for in this study. Measurement of these effects would be necessary to make any claims about the accuracy of the water velocity estimates. ADCP survey line results The data collected by the vessel-mounted A D C P on the J.P. Tully were first examined on small spatial (14 km x 1.5 km) and small temporal (19 h) scales. The backscatter residuals were examined both as in-situ vertical sections, as time series for various loca-tions across the shelf break and slope and as time series for various depth bins along the entire line. The current patterns were examined as in-situ vertical sections, and as time series for various depth bins along the entire line. The zooplankton distributions represented by the backscatter residuals exhibited a definite structure which persisted during the first seven daylight hours of the time series. Chapter 6. Summary and Recommendations 184 This structure consisted of a strong surface scattering layer all across the shelf and slope, a depletion region between depths of 80-150 m over the slope, and a strong bottom scattering layer over the shelf break and slope. Horizontal offshore motion of aggregations of scatterers over the shelf was observed and the coincident water flows confirmed that the motion was due to advection by offshore flow. The persistent structure observed during the daytime changed drastically at sunset. The nighttime structure consisted of near-uniform scattering levels throughout the water column over the shelf and shelf break and a strong surface scattering layer across the shelf and slope. The changes observed at sunset were attributed to vertical migrations of euphausiids from depths of 150-200 m over the shelf break to the surface, with mean swimming speeds of 1-1.5 cm/s. A strong scattering layer was present over the shelf in the last few hours of the time series, and is believed to have been comprised of a high concentration of euphausiids which were advected into the area by a reversal in water flow from southeastward to northwestward. Such an advection was confirmed by the coincident A D C P measurements of the alongshore water flow. This scattering layer underwent an abrupt downwards migration from the surface waters to the near-bottom waters (100-120 m) just prior to sunrise, with mean swimming speeds of 3-4 cm/s. The current patterns were consistent both with the summer mean southeastward-flowing shelf-break current (Freeland et al., 1984; Thomson et al., 1989b) and the tidal flows as calculated by Foreman's barotropic model (Foreman and Walters, 1990). Signif-icant baroclinic tidal effects were observed as periodically occurring vertical shears with 12 hour periods. Absolute water velocities were calculated from bottom-tracking for most of the survey period. The errors in the absolute water velocities from bottom-tracking ship velocities were calculated by assuming uniform flows over horizontal spatial scales of 500 m. The errors were about 5-6 cm/s, which confirmed the estimates of 5 cm/s made previously. The assumption of horizontal uniformity in currents was justified using Chapter 6. Summary and Recommendations 185 estimates of the maximum horizontal shear present over the shelf break. The absolute water velocities from GPS ship velocities were not accurate enough to yield any consistent values, and so were discarded. Coastwide survey results The data collected in June 1993 with the A D C P on the J.P. Tully were also examined over spatial scales of about 100 km and temporal scales of 1-3 weeks. The zooplankton scattering levels had similar day and night cross-shelf structures to those observed in the A D C P line survey. Current patterns were consistent with a summer mean southeastward flowing shelf-break current (Freeland et al., 1984), an estuarine type flow at the mouth of Juan de Fuca Strait (Labrecque et al., 1994), and a counter-clockwise rotating eddy over Juan de Fuca Canyon (Ware and Thomson, 1992). The errors in water velocity components were calculated by assuming horizontal uni-formity in currents over spatial scales of 500 rri, as was done previously for the A D C P survey line data. The errors were found to be 7-8 cm/s when using bottom-tracking to estimate ship velocities, which were consistent (to within 2-3 cm/s) both with the estimates made for the A D C P survey line data (5-6 cm/s) and with the estimates made using the RDI (1989) method (5 cm/s). The errors in water velocity components were about 30 cm/s when using GPS to estimate ship velocities. This estimate was lower than the 40 cm/s estimate based on differences between GPS and bottom-tracking ship veloc-ities. However, the estimate of 30 cm/s is considered more accurate as it does not involve the bottom-tracking errors. The variability in absolute, water, velocities resulting from GPS estimates of ship velocities was found to be large enough to mask any water flow patterns and therefore, only the estimates of absolute velocities using bottom-tracking were used. Spatial averaging of the absolute velocities using bottom-tracking resulted in water velocity errors of about 2-3 cm/s. Chapter 6. Summary and Recommendations 186 Summer mean patterns Averaging the data over the month of June revealed only the gross features of both backscatter residuals and water velocities. The zooplankton distributions observed in the in-situ vertical sections were found to persist all along the coast from the mouth of Juan de Fuca Strait to Estevan Point. The surface layer (6-30 m depth) revealed uniform backscatter levels both day and night across the shelf and slope. The mid-depth (80-100 m) depletion layer over the slope persisted all along the coast. The mid-depths revealed a banded zooplankton structure along the shelf break which reflected the boundary between the depletion layer over the slope and the higher zooplankton concentrations over the shelf. The deep water (175-200 m depth) also revealed a banded structure along the shelf-break, due to the boundary between the slope water zooplankton concentrations and the high concentrations near the bottom at the shelf-break. The banded structures in the waters below 80 m along the shelf-break were more pronounced (higher cross-shelf gradients in scattering levels) during the day. The currents observed in the in situ vertical sections also were found to persist all along the coast from the mouth of Juan de Fuca Strat to Estevan Point. A southeastward flowing surface shelf-break current with speeds of about 20 cm/s was observed along the shelf break from Estevan Point to the waters over the mouth of Juan de Fuca Canyon. Outflow from Juan de Fuca Strait was observed to be strong (30-40 cm/s) in the surface waters (6-30 m) and weak (5-10 cm/s) in the deeper waters (78-102 m). A counter-clockwise rotating eddy was observed over Juan de Fuca Canyon. These currents were consistent with previous observations (Freeland et al., 1984; Ware and Thomson, 1992; Labrecque et al., 1994). Viability of spatial averaging Accurate summer mean zooplankton distributions and currents are difficult to obtain Chapter 6. Summary and Recommendations 187 from vessel-mounted A D C P data due to variability in backscatter residuals and in water velocities at semidiurnal and diurnal frequencies. Zooplankton distributions showed high variability at sunset and sunrise due to migration. Even after separation of the profiles into day and night, the migrations disrupted the mean patterns for a few hours prior to sunset and sunrise. The zooplankton distributions also showed high variability through-out the day due to advection of large patches of zooplankton. Water flow showed high variability at diurnal and semidiurnal frequencies throughout the survey area due to tidal flows. It was found that there was not enough data sampling to well resolve the mean summer zooplankton distributions and currents. To estimate how many sampling periods 1 (for a given location) would be needed to resolve the mean currents and backscatter residuals to within 5 dB and 5 cm/s, respectively, the variances in the mean spatially averaged water velocity components and backscatter residuals were examined for the data taken off Barkley Sound (vertical sections 1-4; see Chapter 5.2.1). The data in sections 1-4 were averaged onto the 22 x 38 grid (Figure 4.16), initializing two grid rows. Each cross-shore location (same grid column) in the two initialized grid rows were then averaged together. The standard deviations on these averages are presented in histogram form (Figure 6.1). The backscatter residuals have significant quantities of data with standard deviations as high as 12 dB. Therefore, the standard errors for one typical cruise with 2-3 sampling periods averaged onto each gridpoint will be 7-9 dB. To reduce this to below 5 dB would require 5-6 sampling periods, which would normally be collected over 2-3 cruises. The currents have significant amounts of data with standard deviations as high as 25 cm/s. Therefore, the standard errors for one, typical cruise with 2-3 sampling periods averaged onto a grid point would be about 15 cm/s. To reduce this to about 5 cm/s would require X A 'sampling period' denotes a period of time during which all ADCP profiles were sequential and averaged onto the same grid point Chapter 6. Summary and Recommendations 188 5 10 (a) Standard deviation on mean backscatter residuals (dB) 15 10 20 30 40 (b) Standard deviation on moan water velocity components (u and v) (cm/s) Figure 6.1: Histograms of standard deviations on mean (a) backscatter residuals and (b) water velocity components (u and v) for the June 1993 cruise. Chapter 6. Summary and Recommendations 189 10-15 sampling periods or 5-8 cruises. The above analysis of sampling needed to resolve mean patterns in A D C P data could be furthered by comparing variances at different locations and at different times of the day to find where and when increased sampling would be needed. A coverage of 10-15 sampling periods per grid-point is not available for any single month for a given year, since there was typically only one cruise per month to the southwest coast of Vancouver Island with the vessel-mounted A D C P . However, such data sampling is available if the data from all cruises during the last seven years are included in the averages. For example, along with the June, 1993 cruise discussed above, there was one cruise in June for each year from 1989-1991 and two cruises in June, 1988, for a total of six June cruises to the southwest coast of Vancouver Island. The mean structures revealed by such an average would be inter-annually persistent. The variability in the zooplankton distributions and current patterns may be higher than that discussed above due to the longer time scales involved. However, such an average would probably reveal the inter-annual mean structures in backscatter residuals and water velocities to within about 2-3 dB and 5 cm/s, respectively. Future work The A D C P survey line conducted in June 1993 yielded some interesting results con-cerning the mean distribution, the diurnal migratory patterns and the advection of zoo-plankton on the Vancouver Island continental margin. The A D C P water velocity data was used primarily to examine the effects of the water flow on the zooplankton distri-butions, but also yielded results concerning the effects of barotropic and baroclinic tidal components on the mean currents. We suggest that a more thorough survey could be completed over the Vancouver Island shelf break as follows: Chapter 6V. Summary and Recommendations 190 Measure the offset angle and scaling bias of the A D C P on the J.P. Tully using one of the methods suggested by Pollard and Read (1989). Set the A D C P bin length to 4 m, the blank-beyond-transmit to 4 m, and the averaging interval to 120 sec. Steam back and forth along a 3-5 km long survey line over the shelf break. The survey line should span the outer, edge (1-2 km) of the shelf and extend a few kilometres seaward of the shelf break. The orientation of the survey line should be perpendic-ular to the mean orientation of the shelf break (200 m isobath). The ship speed should be roughly constant at about. 10 knots, so as to cover the survey line in about two hours. . Steam the survey line for at least a 48 hour period. This would ensure two diurnal and four semidiurnal tidal periods as well as two diurnal plankton migration periods. The ship should be able to steam the survey line about 20 times in 48 hours, yielding enough coverage to give accurate estimates of the mean structures present in the area. • • Take water property data at about five selected sites along the survey line. This should include C T D casts to enable calculations of the internal tidal modes and corre-lations between.the A D C P currents, zooplankton distributions and water mass displacements. Perform net tows at about three locations: Over the shelf; over the shelf break; and over the slope. Net tows at these locations should be done at least once during each day and once during each night. Multiple nets should be used to determine the abundances of zooplankton species at representative depths of 25, 50, 75, 100, 150, 200 and 300 m. This would enable calibration of the backscatter data to yield absolute estimates of biomass throughout the survey line. It would also enable Chapter 6. Summary and Recommendations 191 identification of the species composition of the aggregations of scatterers observed. Preferably, the survey should be completed in the summer when the plankton concentra-tions are highest. Similar surveys could be completed in the winter and spring to observe the seasonal variations in zooplankton distributions. Bibliography Appell, G. F. , Bass, P. D., and Metcalf, M . A . (1991). Acoustic Doppler current profiler performance in near surface and bottom boundaries. Journal of Oceanic Engineering, 16(4):390-396. Becker, R. A . , Chambers, J . M . , and Wilks, A . R. (1988). The New S Language. Wadsworth and Brooks/Cole Advanced Books and Software, Pacific Grove, C A . Bevington, P. R. (1969). Data Reduction and Error Analysis for the Physical Sciences. McGraw-Hill Book Company, New York. Bollens, S. M . , Frost, B . W., and Lin, T. S. (1992). Recruitment, growth and diel vertical migration of euphausia pacifica in a temperate fjord. Marine Biology, 114:219-228. Burd, B . J . and Thomson, R. E. (1993). Flow volume calculations based on three-dimensional current and net orientation data. Deep-Sea Research, 40(6):1141—1153. Burd, B . J . and Thomson, R. E. (1994). Hydrothermal venting at Endeavour Ridge: effect on zooplankton biomass throughout the water column. Deep-Sea Research, 41(9):1407-1423. Burd, B . J . , Thomson, R. E. , and Jamieson, G. S. (1992). Composition of a deep scattering layer overlying a mid-ocean ridge hydrothermal plume. Marine Biology, 113:517-526. Crawford, W. R. and Thomson, R. E. (1982). Continental shelf waves of diurnal period along Vancouver Island. Journal of Geophysical Research, 87:9516-9522. 192 BIBLIOGRAPHY 193 Crawford, W. R. and Thomson, R. E. (1984). Diurnal period continental shelf waves along Vancouver Island: A comparison of observations with theoretical models. Journal of Physical Oceanography, 14(10):1629-1646. Cuypers, L. (1993). Oceanplot 1.01 Manual. Hermes Computing Services, Victoria, B .C . Flagg, C. N . and Smith, S. L. (1989). On the use of the acoustic Doppler current profiler to measure zooplankton abundance. Deep-Sea Research, 36(3):455-474. Flather, R. A . (1987). A tidal model of the northeast Pacific. Atmosphere-Ocean, 25(l):22-45. Flather, R. A . (1988). A numerical model investigation of tides and diurnal period continental shelf waves along Vancouver Island. Journal of Physical Oceanography, 18:115-139. Foreman, M . G. and Freeland, H. J . (1991). A comparison of techniques for tide removal from ship-mounted acoustic Doppler measurements along the southwest coast of Vancouver Island. Journal of Geophysical Research, 96(C9):17007-17021. Foreman, M . G. and Walters, R. A . (1990). A finite-element tidal model for the southwest coast of Vancouver Island. Atmosphere-Ocean, 28(3):281-287. Freeland, H . J . , Crawford, W. R., and Thomson, R. E. (1984). Currents along the Pacific coast of Canada. Atmosphere-Ocean, 22(2):151—172. Freeland, H . J . and Denman, K . L. (1982). A topograhically controlled upwelling center off southern Vancouver Island. Journal of Marine Research, 40:1069-1093. Gargett, A . E . (1988). A 'large eddy' approach to acoustic remote sensing of turbulent kinetic energy dissipation rate e. Atmosphere-Ocean, 26(4):483-508. BIBLIOGRAPHY 194 Genin, A . , Haury, L. , and Greenblatt, P. (1988). Interactions of migrating zooplankton with shallow topography: predation by rockfishes and intensification of patchiness. Deep-Sea Research, 35(2A):151-175. Geyer, W . R. and Signell, R. (1990). Measurements of tidal flow around a headland with a shipboard acoustic Doppler current profiler. Journal of Geophysical Research, 95(C3):3189-3197. Greenlaw, C. F . (1977). Backscattering spectra of preserved zooplankton. Journal of the Acoustical Society of America, 62(l):44-52. Guerin-Ancey, O. and David, P. M . (1993). Use of a multibeam-multifrequency sounder to study the distribution of small zooplankton. Deep-Sea Research, 40(1): 119—128. Heywood, K . J. , Scrope-Howe, S., and Barton, E. D. (1991). Estimation of zooplankton abundance from shipborne A D C P backscatter. Deep-Sea Research, 38(6):677-691. Holliday, D. V . and Pieper, R. E. (1980). Volume scattering strengths and zooplank-ton distributions at acoustic frequencies between 0.5 and 3 MHz. Journal of the Acoustical Society of America, 67(1):135-146. Huyer, A . , Kosro, P. M . , Fleishbein, J . , Ramp, S. R., Stanton, T., Washburn, L. , Chavez, F. P., Cowles, T. J . , Pierce, S. D., and Smith, R. L. (1991). Currents and water masses of the coastal transition zone off northern California, June to August 1988. Journal of Geophysical Research, 96(C8):14809-14831. Joyce, T. M . , Bitterman, Jr., D. S., and Prada, K . E. (1982). Shipboard acoustic profiling of upper ocean currents. Deep-Sea Research, 29(7A):903-913. Kampa, E . M . (1975). Observations of a sonic-scattering layer during the total solar eclipse, 30 June, 1973. Deep-Sea Research, 22(6):417-423. BIBLIOGRAPHY 195 Kaneko, A . , Koterayama, W., Honji, H . , Mizuno, S., Kawatate, K . , and Gordon, R. L. (1990). Cross-stream survey of the upper 400m of the Kuroshio by an A D C P on a towed fish. Deep-Sea Research, 37(5):875-889. Kosro, P., Huyer, A . , Ramp, S. R., Smith, R. L. , Chavez, F. P., Cowles, T. J . , Abbott, M . R., Strub, P. T., Barber, R. T., Jessen, P., and Small, L. F. (1991). The structure of the transition zone between coastal waters and the open ocean off northern Cali-fornia, winter and spring 1987. Journal Geophysical Research, 96(C8):14707-14730. Kosro, P. M . (1985). Shipboard Acoustic Current Profiling During the Coastal Ocean Dynamics Experiment. PhD thesis, University of California, San Diego. Labrecque, A . M . , Thomson, R. E. , Stacey, M . W., and Buckley, J . R. (1994). Residual currents in Juan de Fuca Strait. Atmosphere-Ocean, 32(2):375-394. Mackas, D. L . (1992). Seasonal cycle of zooplankton off southwestern British Columbia. Canadian Journal of Fisheries and Aquatic Sciences, 49:903-921. Mackas, D. L . , Louttit, G. C , and Austin, M . J. (1980). Spatial distribution of zoo-plankton and phytoplankton in British Columbia coastal waters. Canadian Journal of Fisheries and Aquatic Sciences, 37(10):1476-1487. Parsons, T. R., Takahashi, M . , and Hargrave, B . (1984). Biological Oceanographic Pro-cesses. Pergamon Press, 3 r d edition. Pieper, R. E. (1971). A Study of the Relationship Between Zooplankton and High-Frequency Scattering of Underwater Sound. PhD thesis, University of British Columbia. BIBLIOGRAPHY 196 Pollard, R. and Read, J . (1989). A method for calibrating shipmounted acoustic Doppler profilers and the limitations of gyro compasses. Journal of Atmospheric and Oceanic Technology, 6(6):859-865. Ramp, S. R., Jessen, P. F. , Brink, K . H . , Niler, P. P., Daggett, F . L . , and Best, J . S. (1991). The physical structure of cold filaments near Point Arena, California, during June 1987. Journal of Geophysical Research, 96(C8): 14859-14883. RDI (1989). Acoustic Doppler Current Profilers Principles of Operation: A Practical Primer. RD Instruments, San Diego, C A . RDI (1990). Calculating absolute backscatter. Technical Report ADCP-90-04, R D In-struments, San Diego, C A . Richter, K . E. (1985). Acoustic determination of small-scale distributions of individual zooplankters and zooplankton aggregations. Deep-Sea Research, 32(2):163-182. Simard, Y . , de Ladurantaye, R., and Therriault, J.-C. (1986). Aggregation of euphausiids along a coastal shelf in an upwelling environment. Marine Ecology, 32:203-215. Simard, Y . and Mackas, D. L. (1989). Mesoscale aggregations of euphausiid sound scat-tering layers on the continental shelf of Vancouver Island. Canandian Journal of Fisheries and Aquatic Sciences, 46:1238-1249. Stat. Sci. (1991). S-PLUS 3.0 Reference Manual. Statistical Sciences, Inc., Seattle, Washington. Thomson, R. E . , Burd, B . J. , Dolling, A . G., Gordon, R., and Jamieson, G. S. (1992a). The deep scattering layer associated with the Endeavour Ridge hydrothermal plume. Deep-Sea Research, 39(1):55—73. BIBLIOGRAPHY 197 Thomson, R. E. , Gordon, R. L. , and Dolling, A . G. (1991). An intense acoustic scattering layer at the top of a mid-ocean ridge hydrothermal plume. Journal of Geophysical Research, 96(C3):4839-4844. Thomson, R. E. , Gordon, R. L. , and Dymond, J . (1989a). Acoustic Doppler current pro-filer observations of a mid-ocean ridge hydrothermal plume. Journal of Geophysical Research, 94(C4):4709-4720. Thomson, R. E. and Gower, J . F. (1985). A wind-induced mesoscale eddy over the Vancouver Island continental slope. Journal of Geophysical Research, 90(C5):8981-8993. Thomson, R. E . , Healey, M . C , Morris, J . F., and Borstand, G. A . (1992b). Commercial troll fishing vessel distribution off Vancouver Island during July 1988: relation to ob-served physical oceanography. Canadian Journal of Fisheries and Aquatic Sciences, 49(4):820-832. Thomson, R. E . , Hickey, B . M . , and LeBlond, P. H . (1989b). The Vancouver Island Coastal Current: fisheries barrier and conduit. In Beamish, R. and McFarlane, G., editors, Effects of Ocean Variability on Recruitment and an Evaluation of Parameters Used in Stock Assessment Models, volume 108, pages 265-296. Canadian Special Publication of Fisheries and Aquatic Sciences. Trump, C. L . (1986). Estimating absolute current velocities by merging shipboard Doppler current profiler data with Loran-C data. In Proceedings of the IEEE third working Conference on Current Measurements, pages 177-183. Urick, R. J . (1983). Principles of Underwater Sound. McGraw-Hill. BIBLIOGRAPHY 198 Ware, D. M. and Thomson, R. E. (1988). Oceanic factors affecting the distribution and recruitment of west coast fisheries. In Sinclair, editor, Report from the National Workshop on Recruitment., volume 1636, pages 31-64. Canadian Technical Report of Fisheries Aquatic Sciences. Ware, D. M. and Thomson, R. E. (1992). La Perouse Project: Seventh annual progress report 1991. DFO Reports. Wilson, C. and Firing, E. (1992). Sunrise swimmers bias acoustic Doppler current profiles. Deep-Sea Research, 39(5):885-892. Appendix A Derivation of the Biomass Formula This appendix describes how to calculate the biomass per unit volume, W, from the acoustic backscatter per unit volume, Sv. To begin, recall (3.3), which related the total acoustic cross section per unit volume, crv, to Sv: 5„ = 101og(^). (A. l ) For n scatterers per unit volume, each with an acoustic cross section, a, the total acoustic cross section per unit volume is: av = n a. (A-2) If each scatterer has a weight, w, then the total biomass per unit volume is: W = nw, (A.3) Combining (A . l ) , (A.2) and (A.3), the total biomass per unit volume is: W = 4 7 T 1 0 5 \" / 1 0 (£) . (A.4) If there are m different types of scatterers present, each with an acoustic cross section, di, a weight, Wi, and an abundance per unit volume, n,-, (i = 1 to m) then (A.2) and (A.3) become: m crv = ^n.cr,-, (A.5) i = l and: m W = £>.u;;- (A.6) 199 Appendix A Derivation of the Biomass Formula 200 Thus, (A.4) becomes: W = 4 n o W p ^ V (A.7) V L ,=i rii Oi ) The abundances nt- in (A.7) can be re-written in terms of relative abundances, by dividing the top and bottom of (A.7) by the total number of scatterers per unit volume, n, as follows: ^ = 47rl05\"/lof/il4^V, Em xti _ . .=1 t ai but, since n ! = 1 The total biomass per unit volume is given by: (A.8) = X>, (A.9) and ^ = 47rl0 5-/ 1 0 f S 1 > W'' )• (A.11) Appendix B Errors on Absolute Velocity from Bottom Tracking The absolute, earth-referenced water velocities, (u, v), were given by (4.17): / .. \\ sin 9 — cos 9 cos 9 sin 9 \\ (K,)n - VB \\ \\ (K,)x J ( B . l ) where (V^)||= along-ship component of relative water velocity, (VW)M= cross-ship component of relative water velocity, 9 = ship heading from compass (degrees), Vb = bottom speed (cm/s). Equation (B. l ) can be re-written in terms of absolute water velocities as: \\ v) sin 9 — cos 9 cos 9 sin 9 \\ V\\\\ (B.2) where V\\\\ = {VW)\\\\ - VB, V L = (VW)L. The variances in (u, v) due to variances in 9, V\\\\ and V± (ae, avn, and ayL) are found using ((Bevington, 1969), (4-9)): 0~u d_ 89 I -. 2 U y v \\ -, 2 d I u V / j dVn 5 / \\ -• 2 dVL u V V J (B.3) where the square of a matrix is defined here as a matrix of the squares of each row. 201 Appendix B Errors on Absolute Velocity from Bottom Tracking 202 Expanding the first term in (B.3) 2 / V I d_ de V v / J cos 0 sin 0 — sin 0 cos 0 Vjj cos 0 + V± sin 0 V± cos 0 — Vjj sin 0 J (V{{ + KL tan 0) 2 N ^ ( K x - ^ t a n ^ ) 2 , \\ 2 C O S 2 0 cTfl Expanding the second term in (B.3): , 2 d dVtt ( u Ov.. lA { sin 0 — cos 0 cos 0 sin f? sin 20 V U / J C O S 2 0 Expanding the third term in (B.3): d u \\ dV± v )\\ ov± = / sin0 -cos0 \\ ( 0^ cos 0 sin 0 1 cos 20 sm20 However, (B.4) (B.5) (B.6) Ml = (K,)n-H (B.7) V± = (Vw)u (B.8) and so, since the fluctuations in water velocities are uncorrelated to the fluctuations in bottom velocities, the errors in absolute water velocities in the ship's reference frame, (B.9) = °l (B.10) 2 2 , 2 °\"V|| =